WO2019033882A1 - 数据处理方法、装置、系统和计算机可读存储介质 - Google Patents
数据处理方法、装置、系统和计算机可读存储介质 Download PDFInfo
- Publication number
- WO2019033882A1 WO2019033882A1 PCT/CN2018/095740 CN2018095740W WO2019033882A1 WO 2019033882 A1 WO2019033882 A1 WO 2019033882A1 CN 2018095740 W CN2018095740 W CN 2018095740W WO 2019033882 A1 WO2019033882 A1 WO 2019033882A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- inertial navigation
- navigation data
- data
- error
- filtering
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
Definitions
- the present disclosure relates to the field of big data technologies, and in particular, to a data processing method, apparatus, system, and computer readable storage medium.
- Three-dimensional laser scanning technology is a new type of mapping technology developed in recent years.
- 3D laser scanning is mainly for high-precision 3D modeling. Its principle is to use laser ranging, densely record the surface 3D coordinates of the target object, reflectivity and texture information, and perform real 3D recording of the spatial target.
- the laser scanner generally uses the inertial navigation system as an in-vehicle device to scan a real scene, acquire laser point cloud data and inertial navigation data, and combine data processing technology to restore a three-dimensional real scene.
- a data processing method includes: periodically acquiring inertial navigation data of a carrier and scanning laser point cloud data obtained by surrounding environment; filtering inertial navigation data according to the measurement parameter; The inertial navigation data and laser point cloud data construct a three-dimensional scene of the surrounding environment.
- filtering the inertial navigation data according to the measurement parameter comprises: selecting a corresponding filtering model according to the measurement parameter, and filtering the inertial navigation data by using the filtering model.
- different filtering models correspond to different error threshold ranges and reference values; filtering the inertial navigation data by using the filtering model includes: determining an error of the inertial navigation data according to the inertial navigation data and the reference value; and navigating the data according to the inertia The result of the comparison of the error and the error threshold range determines whether the inertial navigation data is filtered.
- filtering the inertial navigation data by using the filtering model includes: determining, according to the difference between the inertial navigation data and the previous period of inertial navigation data, The error of the inertial navigation data, if the error of the inertial navigation data does not satisfy the first threshold range, the inertial navigation data is filtered out; or, the reference value is the inertial navigation data to the first period of inertial navigation data is not filtered
- filtering the inertial navigation data by using the filtering model includes: from the inertial navigation data to the first period of inertial navigation data, respectively calculating the inertial navigation data of each adjacent two cycles that are not filtered out.
- filtering the inertial navigation data by using the filtering model includes: determining the error of the inertial navigation data according to the difference between the inertial navigation data and the average value of the inertial navigation data of each period, and the error of the inertial navigation data is not In the case where the third threshold range is satisfied, the inertial navigation data is filtered out.
- the inertial navigation data is filtered using the filtering model in the case where the reference value includes the inertial navigation data of the previous period of the inertial navigation data, the error rate of the inertial navigation data, and the error rate of the inertial navigation data of the previous period.
- the method comprises: determining an error of the inertial navigation data according to a difference between the compensation value of the inertial navigation data and the compensation value of the inertial navigation data of the previous period, and if the error of the inertial navigation data does not satisfy the fourth threshold range, the inertial navigation is performed.
- the data is filtered out; or, in the case where the reference value includes the inertial navigation data to the inertial navigation data of each period that is not filtered out in the first period of inertial navigation data and the error rate of the inertial navigation data of each period, the inertia of the filter model is used.
- the filtering of the navigation data includes: calculating inertia navigation data from the inertial navigation data to the first period of inertial navigation data, respectively calculating the difference of the compensation values of the adjacent two-cycle inertial navigation data that are not filtered, and determining the inertia according to the average value of the respective differences
- the error of the navigation data, the error in the inertial navigation data does not meet the fifth threshold
- the reference value includes the inertial navigation data to the first period of inertial navigation data that is not filtered out of the periodic inertial navigation data and the error rate of the inertial navigation data of each cycle
- filtering the inertial navigation data by using the filtering model includes: calculating an average value of the inertial navigation data to the inertia navigation data of the first period of inertial navigation data that is not filtered out, and the compensation value according to the inertial navigation data.
- the difference from the average value determines the error of the inertial navigation data. If the error of the inertial navigation data does not satisfy the sixth threshold range, the inertial navigation data is filtered out; wherein the compensation value of the inertial navigation data is inertial navigation data. The product of the corresponding error rate.
- the method further comprises: determining a reference error based on an average of the errors of the filtered inertial navigation data and a corresponding reference error threshold, and correcting the filtered inertial navigation data by using the reference error;
- the reference error threshold corresponds to the filtering model.
- the method before constructing the three-dimensional scene according to the filtered inertial navigation data and the laser point cloud data, the method further includes: deleting laser point cloud data corresponding to the laser beam that does not meet the first preset angle range; or The laser point cloud data corresponding to the laser line in the laser beam of the second predetermined angle range is deleted.
- constructing the surrounding three-dimensional scene according to the filtered inertial navigation data and the laser point cloud data comprises: converting the laser point cloud data acquired in each cycle from the laser coordinate system to the inertia acquired at the same time and not filtered.
- the geographic coordinate system of the navigation data; the laser point cloud data of each cycle in the geographic coordinate system is superimposed to construct a three-dimensional scene of the surrounding environment.
- the inertial navigation data includes at least one of motion pose data and geographic location data
- the motion pose data includes at least one of heading data, pitch data, and roll data
- the geographic location data includes longitude, latitude, and elevation.
- the geographic location data is obtained from the satellite navigation system and corrected by the inertial navigation system according to the direction of inertia, the motion situation, and the force of each degree of freedom.
- the measurement parameters include at least one of a road condition, a carrier speed, an acquisition frequency of inertial navigation data, a manufacturer of the inertial navigation system, a model of the inertial navigation system, and an inertial navigation system hardware error.
- a data processing apparatus includes: a data acquisition unit configured to periodically acquire inertial navigation data of a carrier and scan laser point cloud data obtained by surrounding environment; and a data filtering unit The method is configured to filter the inertial navigation data according to the measurement parameter; the modeling unit is configured to construct a three-dimensional scene of the surrounding environment according to the filtered inertial navigation data and the laser point cloud data.
- the data filtering unit is configured to select a corresponding filtering model based on the measurement parameters, and filter the inertial navigation data using the filtering model.
- the different filtering models correspond to different error threshold ranges and reference values; the data filtering unit is configured to determine an error of the inertial navigation data based on the inertial navigation data and the reference value, based on the error and error threshold of the inertial navigation data The result of the comparison of the ranges determines whether the inertial navigation data is filtered.
- the data filtering unit is configured to determine from the difference between the inertial navigation data and the adjacent previous periodic inertial navigation data. The error of the inertial navigation data, if the error of the inertial navigation data does not satisfy the first threshold range, the inertial navigation data is filtered out; or, the reference value is the inertial navigation data to the first period of inertial navigation data is not filtered.
- the data filtering unit is configured to calculate the difference between the inertial navigation data of each adjacent two periods that is not filtered out from the inertial navigation data to the first period of inertial navigation data, and according to The average value of each difference is used to determine the error of the inertial navigation data, and the inertial navigation data is filtered out if the error of the inertial navigation data does not satisfy the second threshold range; or, the reference value is inertial navigation data to the first period In the case of the case of the
- the data filtering unit is configured to navigate according to inertia The difference between the compensation value of the data and the compensation value of the inertial navigation data of the previous period determines the error of the inertial navigation data.
- the inertial navigation data filtering unit is configured to navigate data from inertia To the inertial navigation data of the first period, respectively calculate the difference of the compensation values of the inertial navigation data of each adjacent two periods that are not filtered out, and determine the error of the inertial navigation data according to the average value of the respective differences, in the inertial navigation data If the error does not satisfy the fifth threshold range, the inertial navigation data is filtered out; or
- the reference value includes the inertial navigation data to the inertial navigation data of each period of the first period of inertial navigation data that is not filtered out and the error rate of the inertial navigation data of each period, the data filtering unit is configured to calculate the inertial navigation data
- the apparatus further includes: a data correction unit configured to determine a reference error based on an average of errors of the filtered inertial navigation data and a corresponding reference error threshold, filtered by the reference error correction Each of the inertial navigation data; wherein the reference error threshold corresponds to the filtering model.
- the apparatus further includes: a laser data screening unit configured to delete the laser point cloud data corresponding to the laser beam that does not meet the first predetermined angular range, or to not meet the second predetermined angular range The laser point cloud data corresponding to the laser line in the laser beam is deleted.
- the modeling unit is configured to convert the laser point cloud data acquired in each cycle from a laser coordinate system to a geographic location coordinate system of the inertial navigation data acquired simultaneously and unfiltered, which will be converted into a geographic location.
- the laser point cloud data of each cycle of the coordinate system is superimposed to construct a three-dimensional scene of the surrounding environment.
- the inertial navigation data includes at least one of motion pose data and geographic location data
- the motion pose data includes at least one of heading data, pitch data, and roll data
- the geographic location data includes longitude, latitude, and elevation.
- the geographic location data is obtained from the satellite navigation system and corrected by the inertial navigation system according to the direction of inertia, the motion situation, and the force of each degree of freedom.
- the measurement parameters include at least one of a road condition, a carrier speed, an acquisition frequency of inertial navigation data, a manufacturer of the inertial navigation system, a model of the inertial navigation system, and an inertial navigation system hardware error.
- a data processing apparatus comprising: a memory; and a processor coupled to the memory, the processor configured to execute any of the foregoing embodiments based on instructions stored in the memory device Data processing method.
- a computer readable storage medium having stored thereon a computer program, wherein the program, when executed by a processor, implements the data processing method of any of the foregoing embodiments.
- a data processing system comprising: the data processing apparatus of any of the foregoing embodiments; and an inertial navigation system configured to acquire inertial navigation data and transmit to the data processing apparatus; the laser scanner , configured to collect laser point cloud data and send it to a data processing device.
- the system further comprises: a satellite navigation system configured to acquire geographic location data and transmit to the inertial navigation system; the inertial navigation system is further configured to be subject to inertial direction, motion conditions, and various degrees of freedom directions The force situation corrects the geographic location data.
- inertial navigation data is filtered according to different measurement parameters, and inertial navigation data that does not conform to the corresponding measurement parameter law can be filtered.
- the present disclosure makes the three-dimensional scene constructed by the laser point cloud data and the inertial navigation data more accurate under various measurement parameters.
- FIG. 1 shows a flow diagram of a data processing method of some embodiments of the present disclosure.
- FIG. 2 is a flow chart showing a data processing method of further embodiments of the present disclosure.
- FIG. 3 illustrates a flow diagram of a data processing method of still other embodiments of the present disclosure.
- FIG. 4 is a flow chart showing a data processing method of still another embodiment of the present disclosure.
- FIG. 5 shows a block diagram of a data processing apparatus of some embodiments of the present disclosure.
- FIG. 6 is a block diagram showing the structure of a data processing apparatus according to further embodiments of the present disclosure.
- FIG. 7 is a block diagram showing the structure of a data processing apparatus according to still another embodiment of the present disclosure.
- FIG. 8 is a block diagram showing the structure of a data processing apparatus of still another embodiment of the present disclosure.
- FIG. 9 is a block diagram showing the structure of a data processing system of still other embodiments of the present disclosure.
- FIG. 1 is a flow chart of some embodiments of a data processing method of the present disclosure. As shown in FIG. 1, the method of this embodiment includes: steps S102 to S106.
- step S102 the inertial navigation data of the carrier and the laser point cloud data obtained by scanning the surrounding environment are periodically acquired.
- the inertial navigation system and the laser scanning scanner are usually disposed on the same carrier (for example, a vehicle). With the movement of the carrier, the inertial navigation system and the laser scanning scanner respectively collect the carrier inertial navigation data and the laser point cloud data of the surrounding environment, and subsequently Combining the two types of data allows you to build a three-dimensional scene of your surroundings.
- the acquisition frequency of the inertial navigation data is, for example, 100 Hz.
- the inertial navigation data includes at least one of motion pose data and geographic location data, and the motion pose data includes at least one of heading data, pitch data, and roll data, and the geographic location data includes at least one of longitude, latitude, and elevation.
- the geographic location data may be geographic location data acquired by a satellite navigation system, such as a GPS (Global Positioning System), and corrected by an inertial navigation system.
- the inertial navigation system corrects the geographic location data collected by the satellite navigation system according to the direction of inertia, the motion situation and the force of each degree of freedom.
- the inertial navigation system When the environment has adverse effects on satellite reception, such as trees or buildings, the inertial navigation system has a built-in gyroscope that can measure the inertial direction of the motion, and then calculate the position at which the next moment will continue to move according to this inertia.
- the inertial navigation system Responsible for solving the estimated coordinates of the arrival position. If there is no inertial navigation system, it can be assisted in the case of a single GPS when the signal is not good.
- the inertial navigation system may also have a built-in accelerometer, which can calculate the force in the direction of multiple degrees of freedom in the space, and correct the original inertial position estimation structure obtained by the gyroscope according to the posture change caused by the force and the magnitude of the force.
- the inertial navigation system will replace the GPS position with the above estimated position.
- the calculation error of the dead reckoning according to the gyroscope will be reset to zero. The above process is much improved compared to the positioning accuracy using a single GPS.
- the laser scanner emits multiple laser beams.
- the emission angles of the multiple lasers are different.
- Each laser beam is composed of a plurality of laser lines, for example, 32 lines, and each line corresponds to a different emission angle in the same vertical plane.
- step S104 the inertial navigation data is filtered according to the measurement parameters.
- the measurement parameters include, for example, at least one of road surface condition, carrier speed, acquisition frequency of inertial navigation data, manufacturer of the inertial navigation system, model of the inertial navigation system, and hardware error of the inertial navigation system.
- Different road conditions such as bumps, can cause the inertia of the carrier to change, resulting in deviations in the acquisition of inertial navigation data.
- Differences in inertial navigation devices can also result in different deviations in the acquisition of inertial navigation data.
- the inertial navigation data will exhibit different variation rules. Therefore, different filtering models can be established for different measurement parameters according to actual conditions, and the filtering model is selected according to the actual measurement parameters to filter the inertial navigation data.
- step S104 includes steps S202 to S204.
- step S202 an error of the inertial navigation data is determined based on the inertial navigation data and the reference value.
- step S204 it is determined whether to filter the inertial navigation data based on the result of the comparison of the error of the inertial navigation data with the error threshold range.
- step S106 a three-dimensional scene is constructed based on the filtered inertial navigation data and the laser point cloud data.
- step S106 includes steps S302 to S304.
- step S302 the laser point cloud data acquired in each cycle is transferred from the laser coordinate system to the unfiltered inertial navigation data geographic coordinate system acquired at the same time.
- step S304 the laser point cloud data converted to the respective periods of the geographic location coordinate system is superimposed to construct a three-dimensional scene of the surrounding environment.
- the laser point cloud is acquired by the laser scanner, and the original three-dimensional scan data is the independent laser coordinate system data with the scanner as the coordinate origin.
- the scanner When the scanner generates motion in the real world, the laser point cloud data superimposed in multiple cycles at this time will be corresponding to the same origin, but not at the same origin after the actual movement, resulting in the complete result cannot be recognized after the data is superimposed. Therefore, the scanner is rigidly connected to the inertial navigation system.
- the inertial navigation system can know the posture and position of the motion, and the laser point cloud data of each cycle is corresponding to the motion posture and position at the same time.
- multi-cycle laser point cloud data can be superimposed to obtain stereoscopic complete 3D laser point cloud data.
- the laser point cloud data processing process can be converted by several coordinate systems. First, the laser point cloud data is transferred from the laser coordinate system to the Cartesian coordinate system, then to the carrier coordinate system, and finally converted into the GPS coordinate system in the carrier coordinate system. .
- the laser point cloud data collected simultaneously with the inertial navigation data can be eliminated, so that the laser point cloud data does not participate in the final modeling.
- the method of the above embodiment is based on different measurement parameters, the inertial navigation data is filtered, and the inertial navigation data that does not conform to the corresponding measurement parameter law can be filtered.
- the method of the above embodiment makes the three-dimensional scene constructed by the laser point cloud data and the inertial navigation data more accurate under various measurement parameters.
- different filtering models correspond to different measurement parameters, reflecting different variations of inertial navigation data, and the inertial navigation data may be filtered according to the filtering model.
- the present disclosure provides some embodiments for filtering inertial navigation data using a filtering model.
- the reference value is the previous period of inertial navigation data of the inertial navigation data.
- the error threshold range corresponding to the selected filter model is, for example, a first threshold range, and the first threshold range may represent a range of normal differences between the two acquired inertial navigation data.
- the first threshold range corresponding to different types of inertial navigation data may include one or more different threshold ranges, including, for example, a heading threshold range, a pitch threshold range, a roll threshold range, a longitude threshold range, a latitude threshold range, and an elevation threshold range. At least one.
- the error of the inertial navigation data is determined according to the difference between the inertial navigation data and the inertial navigation data of the previous period. If the error of the inertial navigation data does not satisfy the first threshold range, the inertial navigation data is filtered out. For example, the difference between the inertial navigation data and the previous inertial navigation data is calculated as the error of the inertial navigation data, and if the error of the inertial navigation data does not satisfy the first threshold range, the inertial navigation data is filtered out.
- the error threshold of the adjacent two collected headings corresponding to the filtering model ranges from 0.01 degrees to 0.05 degrees, and the difference between each adjacent two measured headings is calculated separately, and the adjacent two are The next time the inertial navigation data does not satisfy the rejection of the error threshold range.
- the inertial navigation data may be sequentially filtered backward from the second measured inertial navigation data, or may be filtered while collecting the inertial navigation data.
- the inertial navigation data measured for the first time may be compared with an initial threshold to determine whether to filter, the initial threshold corresponding to the filtering model, that is, the initial threshold may be different under different measurement parameters.
- the filtered inertial navigation data can no longer participate in the calculation, that is, the reference value is the previous period of inertial navigation data that is not filtered out. If the data of the first N consecutive cycles adjacent to the inertial navigation data is filtered out, N is greater than the pre- With the value set, an alarm can be issued to stop the measurement.
- the reference value is inertial navigation data to each period of inertial navigation data that is not filtered out in the first period of inertial navigation data
- the selected filter model corresponds to a second threshold range
- the second threshold range represents, for example, each The normal range of the average difference between the inertial navigation data acquired twice adjacent.
- the second threshold range corresponding to different types of inertial navigation data may include one or more different types of threshold ranges, including, for example, a heading threshold range, a pitch threshold range, a roll threshold range, a longitude threshold range, a latitude threshold range, and an elevation threshold range. At least one of them.
- the inertial navigation data is filtered out.
- Each difference can be averaged as an error in the inertial navigation data.
- n+1th measured inertial navigation data value n+1 where n is a positive integer greater than or equal to 1, and value n+1 may represent pitch data
- the error of value n+1 may be expressed as i is a positive integer, assuming n n unfiltered inertial navigation data before value n+1 . If X n+1 does not satisfy the second threshold range, then value n+1 is filtered out.
- the reference value is an average of the inertial navigation data to the periodic inertial navigation data that is not filtered out of the first periodic inertial navigation data.
- the error threshold range corresponding to the selected filter model is, for example, a third threshold range, and the third threshold range may represent a normal difference range of the inertial navigation data and the previous inertial navigation data average.
- the third threshold range corresponding to different types of inertial navigation data may include one or more different threshold ranges, including, for example, a heading threshold range, a pitch threshold range, a roll threshold range, a longitude threshold range, a latitude threshold range, and an elevation threshold range. At least one.
- the error of the inertial navigation data is determined according to the difference between the inertial navigation data and the average value of the inertial navigation data of each period, and the inertial navigation data is filtered out if the error of the inertial navigation data does not satisfy the third threshold range.
- the difference between the inertial navigation data and the average of the periodic inertial navigation data is determined as the error of the inertial navigation data.
- the reference value may also be an average value of the inertial navigation data of each period that is not filtered out from the inertial navigation data to the M-th cycle inertial navigation data, and M is a preset value.
- the reference value includes inertial navigation data for the previous period of the inertial navigation data, an error rate for the inertial navigation data, and an error rate for the inertial navigation data of the previous period.
- the error threshold range corresponding to the selected filter model is, for example, a fourth threshold range, and the fourth threshold range may represent a range of normal differences between the compensation values of the adjacent two acquired inertial navigation data. The error caused by the environment or the device itself can be compensated to some extent, and the error of the compensated inertial navigation data is filtered if it cannot satisfy the error threshold range.
- the inertial navigation data compensation value may be the product of the inertial navigation data and the corresponding error rate, and the error rate varies according to the difference of the filtering model and the measurement period.
- the fourth threshold range corresponding to different types of inertial navigation data may include one or more different threshold ranges, including, for example, a heading threshold range, a pitch threshold range, a roll threshold range, a longitude threshold range, a latitude threshold range, and an elevation threshold range. At least one.
- the error of the inertial navigation data is determined according to the difference between the compensation value of the inertial navigation data and the compensation value of the inertial navigation data of the previous period. If the error of the inertial navigation data does not satisfy the fourth threshold range, the inertial navigation is performed. The data is filtered out. For example, the inertial navigation data and the adjacent previously measured inertial navigation data are respectively multiplied by the corresponding error rate to obtain a difference as an error of the inertial navigation data.
- the reference value includes inertial navigation data to each period of inertial navigation data that is not filtered out in the first periodic inertial navigation data, and an error rate of inertial navigation data for each period.
- the error threshold range corresponding to the selected filter model is, for example, a fifth threshold range, and the fifth threshold range may represent a normal average difference range between the compensation values of the adjacent two acquired inertial navigation data.
- the inertial navigation data compensation value may be the product of the inertial navigation data and the corresponding error rate, and the error rate varies according to the difference of the filtering model and the measurement period.
- the fifth threshold range corresponding to different types of inertial navigation data may include one or more different threshold ranges, including, for example, a heading threshold range, a pitch threshold range, a roll threshold range, a longitude threshold range, a latitude threshold range, and an elevation threshold range. At least one.
- the inertial navigation data to the first period of inertial navigation data respectively calculate the difference of the compensation values of the adjacent two periods of inertial navigation data that are not filtered out, and determine the error of the inertial navigation data according to the average value of the respective differences If the error of the inertial navigation data does not satisfy the fifth threshold range, the inertial navigation data is filtered out. For example, the average of the respective differences is determined as the error of the inertial navigation data.
- n+1th measured inertial navigation data value n+1 where n is a positive integer greater than or equal to 1, and value n+1 may represent pitch data
- the error of value n+1 may be expressed as Where p i and p i-1 represent the error rates of the inertial navigation data measured at the ith and i-1th, respectively, assuming that there are n unfiltered inertial navigation data before value n+1 . If X n+1 does not satisfy the fifth threshold range, then value n+1 is filtered out.
- the mean value of the absolute values of the respective differences after taking the absolute value of the difference between the adjacent two cycles of the inertial navigation data as the error of the inertial navigation data.
- the error of value n+1 can be expressed as
- the reference value includes inertial navigation data to each cycle of inertial navigation data that is not filtered out of the first periodic inertial navigation data, and an error rate of inertial navigation data for each cycle.
- the error threshold range corresponding to the selected filter model is, for example, a sixth threshold range, and the sixth threshold range may represent a normal difference range of the inertial navigation data compensation value and the average value of the compensation values of the previous inertial navigation data.
- the inertial navigation data compensation value may be the product of the inertial navigation data and the corresponding error rate, and the error rate varies according to the difference of the filtering model and the measurement period.
- the sixth threshold range corresponding to different types of inertial navigation data may include one or more different threshold ranges, including, for example, a heading threshold range, a pitch threshold range, a roll threshold range, a longitude threshold range, a latitude threshold range, and an elevation threshold range. At least one.
- n+1th measured inertial navigation data value n+1 where n is a positive integer greater than or equal to 1, and value n+1 may represent pitch data
- the error of value n+1 may be expressed as Where p i and p i-1 represent the error rates of the inertial navigation data measured at the ith and i-1th, respectively, assuming that there are n unfiltered inertial navigation data before value n+1 . If X n+1 does not satisfy the sixth threshold range, then value n+1 is filtered out.
- the reference value may also be an average value of the inertial navigation data compensation values of the inertial navigation data to the M-cycle inertial navigation data that are not filtered out, and M is a preset value.
- the different filtering models correspond to different sets of reference error thresholds
- the reference error thresholds represent adjustment thresholds of the average values of the filtered errors of the respective inertial navigation data, ie, the errors of the filtered individual inertial navigation data may be Accept the scope.
- the reference error threshold group corresponding to different inertial navigation data may include one or more different reference error thresholds, including, for example, heading reference error threshold, pitch reference error threshold, roll reference error threshold, longitude reference error threshold, latitude reference error threshold At least one of the elevation reference error thresholds.
- the reference error is determined according to the average value of the filtered error of each inertial navigation data and the corresponding reference error threshold, for example, calculating the average value of the filtered error of each inertial navigation data, and calculating the difference between the average value and the corresponding reference error As the reference error; each of the inertial navigation data filtered by the reference error correction, that is, the filtered inertial navigation data is respectively subtracted from the reference error.
- the difference between the average value of the error of each filtered inertial navigation data and the reference error threshold is e, and the n+1th measured inertial navigation data value n+1 is filtered out, then the correction of value n+1 The value is value n+1 -e.
- the modified inertial navigation data can be reconfigured to model the data to make up for the lack of data.
- the first embodiment is used as a rough filtering of inertial navigation data, and further filtering is performed by other embodiments.
- step 306 is a flow chart of still another embodiment of a data processing method of the present disclosure. As shown in FIG. 4, before step 306, the method may further include: steps S402-S408.
- step S402 the inertial navigation data and the laser point cloud data are periodically acquired.
- step S404 the inertial navigation data is filtered according to the measurement parameters.
- step S406 the laser point cloud data corresponding to the laser beam or the laser line is deleted according to the test demand angle.
- the laser point cloud data corresponding to the laser beam that does not meet the first preset angle range is deleted; or the laser point cloud data corresponding to the laser line in the laser beam that does not meet the second preset angle range is delete.
- the laser scanner can emit multiple lasers in a 360-degree angular range parallel to the ground, and each laser can be divided into multiple laser lines in a plane perpendicular to the ground. Therefore, part of the laser beam or laser line can be eliminated according to actual test requirements. For example, it is only necessary to measure an object in the ground range, and the point cloud data corresponding to the laser line obliquely upward can be eliminated. It is only necessary to measure the scene on the left side of the vehicle's forward direction, and the laser beam on the right side of the vehicle's forward direction can be removed. This can save data storage and calculations and improve efficiency.
- Step S406 may be performed in parallel with step S404 without prioritization.
- step S408 a three-dimensional scene is constructed based on the filtered inertial navigation data and the laser point cloud data.
- Steps S402, 404, and 408 can refer to the embodiments corresponding to steps S102, 104, and 108 in the foregoing embodiment, respectively.
- the present disclosure also provides a data processing apparatus, which will be described below in conjunction with FIG.
- FIG. 5 is a structural diagram of some embodiments of a data processing apparatus according to the present disclosure. As shown in FIG. 5, the apparatus 50 includes a data acquisition unit 502, a data filtering unit 504, and a modeling unit 506.
- the data acquisition unit 502 is configured to periodically acquire inertial navigation data of the carrier and scan the laser point cloud data obtained by the surrounding environment.
- the data acquisition unit 502 can perform the method of step S102.
- the inertial navigation data includes at least one of motion pose data and geographic location data, and the motion pose data includes at least one of heading data, pitch data, and roll data, and the geographic location data includes at least one of longitude, latitude, and elevation;
- the data is obtained from the satellite navigation system and corrected by the inertial navigation system according to the direction of inertia, the motion situation and the force of each degree of freedom.
- the data filtering unit 504 is configured to filter the inertial navigation data according to the measurement parameters.
- the data filtering unit 504 can perform the method of step S104.
- the measurement parameters include, for example, at least one of road surface condition, carrier speed, acquisition frequency of inertial navigation data, manufacturer of the inertial navigation system, model of the inertial navigation system, and hardware error of the inertial navigation system.
- the data filtering unit 504 is configured to select a corresponding filtering model based on the measurement parameters, and to filter the inertial navigation data using the filtering model.
- different filtering models correspond to different error threshold ranges and reference values.
- the data filtering unit 504 is configured to determine an error of the inertial navigation data based on the inertial navigation data and the reference value, and determine whether to filter the inertial navigation data according to the comparison result of the error of the inertial navigation data and the error threshold range.
- the data filtering unit 504 is configured to calculate a difference between the inertial navigation data and the adjacent previous periodic inertial navigation data, The error of the inertial navigation data is determined. If the error of the inertial navigation data does not satisfy the first threshold range, the inertial navigation data is filtered out.
- the data filtering unit 504 is configured to navigate from the inertial navigation data to the first cycle. Inertial navigation data, respectively calculating the difference of each adjacent two inertial navigation data that is not filtered, and determining the error of the inertial navigation data according to the average value of each difference, and the error of the inertial navigation data does not satisfy the second threshold In the case of the range, the inertial navigation data is filtered out.
- the data filtering unit 504 is configured to The difference between the average values of the inertial navigation data of each period determines the error of the inertial navigation data, and the inertial navigation data is filtered out if the error of the inertial navigation data does not satisfy the third threshold range.
- the data filtering unit 504 is configured in the case where the reference value includes the inertial navigation data adjacent to the previous period of inertial navigation data, the error rate of the inertial navigation data, and the error rate of the adjacent previous period of inertial navigation data.
- the inertia is The navigation data is filtered out.
- the data filtering unit 504 is The method is configured to calculate the difference of the compensation values of the inertial navigation data of each adjacent two periods that are not filtered out from the inertial navigation data to the inertial navigation data of the first period, and determine the error of the inertial navigation data according to the average value of the respective differences. If the error of the inertial navigation data does not satisfy the fifth threshold range, the inertial navigation data is filtered out.
- the data filtering unit 504 is The method is configured to calculate an average value of the inertial navigation data to the inertia navigation data of the first period of inertial navigation data that is not filtered out, and the average value of the compensation value of the inertial navigation data and the compensation value of the inertial navigation data of each period. The difference is determined by the error of the inertial navigation data. If the error of the inertial navigation data does not satisfy the sixth threshold range, the inertial navigation data is filtered out. The compensation value of the inertial navigation data is the product of the inertial navigation data and the corresponding error rate.
- the modeling unit 506 is configured to construct a three-dimensional scene based on the filtered inertial navigation data and the laser point cloud data.
- the modeling unit 506 can perform the method of step S106.
- the modeling unit 506 is configured to convert the laser point cloud data acquired in each cycle from a laser coordinate system to a geographic location coordinate system of the inertial navigation data acquired simultaneously and unfiltered, which is converted into The laser point cloud data of each cycle of the geographic coordinate system is superimposed to construct a three-dimensional scene of the surrounding environment.
- FIG. 6 is a structural diagram of some embodiments of a data processing apparatus according to the present disclosure.
- the device 60 includes: a data acquiring unit 602, a data filtering unit 604, and a modeling unit 606.
- the three units respectively implement the data acquiring unit 502, the data filtering unit 504, and the modeling unit 506 in FIG. Similar features.
- the data correction unit 608 is configured to determine the reference error according to the average value of the filtered error of each of the inertial navigation data and the corresponding reference error threshold, and correct the filtered inertial navigation data by using the reference error.
- the reference error threshold corresponds to the filtering model.
- the data processing device 60 may further include: a laser data screening unit 610 configured to delete laser point cloud data corresponding to the laser beam that does not meet the first preset angle range, or to not meet the first preset angle range The laser point cloud data corresponding to the laser beam is deleted, or the laser point cloud data corresponding to the laser line in the laser beam that does not meet the second preset angle range is deleted.
- a laser data screening unit 610 configured to delete laser point cloud data corresponding to the laser beam that does not meet the first preset angle range, or to not meet the first preset angle range The laser point cloud data corresponding to the laser beam is deleted, or the laser point cloud data corresponding to the laser line in the laser beam that does not meet the second preset angle range is deleted.
- the data processing apparatus in the embodiments of the present disclosure may each be implemented by various computing devices or computer systems, as described below in connection with FIGS. 7 and 8.
- Figure 7 is a block diagram of some embodiments of a data processing apparatus of the present disclosure.
- the apparatus 70 of this embodiment includes a memory 710 and a processor 720 coupled to the memory 710, the processor 720 being configured to perform any of the implementations of the present disclosure based on instructions stored in the memory 710.
- the data processing method in the example is a block diagram of some embodiments of a data processing apparatus of the present disclosure.
- the apparatus 70 of this embodiment includes a memory 710 and a processor 720 coupled to the memory 710, the processor 720 being configured to perform any of the implementations of the present disclosure based on instructions stored in the memory 710.
- the data processing method in the example is a block diagram of some embodiments of a data processing apparatus of the present disclosure.
- the memory 710 may include, for example, a system memory, a fixed non-volatile storage medium, or the like.
- the system memory stores, for example, an operating system, an application, a boot loader, a database, and other programs.
- FIG. 8 is a block diagram of another embodiment of a data processing apparatus of the present disclosure.
- the apparatus 80 of this embodiment includes a memory 810 and a processor 820, which are similar to the memory 710 and the processor 720 of FIG. 7, respectively.
- Input and output interface 830, network interface 840, storage interface 850, and the like can also be included. These interfaces 830, 840, 850 and the memory 810 and the processor 820 can be connected, for example, via a bus 860.
- the input/output interface 830 provides a connection interface for input and output devices such as a display, a mouse, a keyboard, and a touch screen.
- the network interface 840 provides a connection interface for various networked devices, such as a database server or a cloud storage server.
- the storage interface 850 provides a connection interface for an external storage device such as an SD card or a USB flash drive.
- the present disclosure also provides a data processing system, which is described below in conjunction with FIG.
- the system 9 is a block diagram of some embodiments of a data processing system of the present disclosure. As shown in FIG. 9, the system 9 includes the data processing apparatus 50, 60, 70 or 80 of any of the foregoing embodiments, and an inertial navigation system 92, a laser scanner 94.
- the inertial navigation system 92 is configured to acquire inertial navigation data and transmit it to the data processing device.
- a laser scanner 94 is configured to acquire laser point cloud data and transmit it to a data processing device.
- the system 9 can also include a satellite navigation system 96 configured to collect geographic location data and transmit it to the inertial navigation system 92.
- a satellite navigation system 96 configured to collect geographic location data and transmit it to the inertial navigation system 92.
- the inertial navigation system 92 is further configured to modify the geographic location data based on the direction of inertia, the motion conditions, and the force conditions in the various degrees of freedom.
- the present disclosure also provides a computer readable storage medium having stored thereon a computer program that, when executed by a processor, implements the steps of the data processing method of any of the foregoing embodiments.
- the data processing method, apparatus, and system of the present disclosure may be configured as an unmanned field in addition to being configured to perform surveying work, and the driverless car may be in accordance with the data processing method and apparatus of the present disclosure.
- a three-dimensional map built by the system to identify the driving environment for unmanned driving may also be configured as an unmanned storage area, and the unmanned storage truck can complete the handling of the goods through the constructed three-dimensional warehouse map.
- embodiments of the present disclosure can be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment, or a combination of software and hardware aspects. Moreover, the present disclosure may take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code. .
- the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
- the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
- These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
- the instructions provide steps that are configured to implement the functions specified in one or more blocks of the flowchart or in a block or blocks of the flowchart.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Remote Sensing (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Navigation (AREA)
Abstract
本公开涉及一种数据处理方法、装置、系统和计算机可读存储介质，涉及大数据技术领域。本公开的方法包括：周期性获取载体的惯性导航数据和扫描周围环境得到的激光点云数据；根据测量参数对惯性导航数据进行过滤；根据过滤后的惯性导航数据与激光点云数据构建周围环境的三维场景。本公开中根据不同的测量参数，惯性导航数据被过滤，不符合对应的测量参数规律的惯性导航数据能够被过滤。本公开使得在各种测量参数下激光点云数据与惯性导航数据配合构建出的三维场景更加准确。
Description
相关申请的交叉引用
本申请是以CN申请号为201710700605.5，申请日为2017年8月16日的申请为基础，并主张其优先权，该CN申请的公开内容在此作为整体引入本申请中。
本公开涉及大数据技术领域，特别涉及一种数据处理方法、装置、系统和计算机可读存储介质。
三维激光扫描技术是近几年发展起来的一种新型的测绘技术。三维激光扫描主要面向高精度的三维建模，其原理是利用激光测距，密集的记录目标物体的表面三维坐标，反射率和纹理信息，对空间目标进行真实的三维记录。
目前，激光扫描仪一般与惯性导航系统一起作为车载设备对现实场景进行扫描，获取激光点云数据以及惯性导航数据，结合数据处理技术，还原三维的现实场景。
发明内容
发明人发现：惯性导航系统会由于设备、路况等的变化导致采集的惯性导航数据不准确，进而使得构建的三维场景出现偏差。
根据本公开的一些实施例，提供的一种数据处理方法，包括：周期性获取载体的惯性导航数据和扫描周围环境得到的激光点云数据；根据测量参数对惯性导航数据进行过滤；根据过滤后的惯性导航数据与激光点云数据构建周围环境的三维场景。
在一些实施例中，根据测量参数对惯性导航数据进行过滤包括：根据测量参数选取对应的过滤模型，利用过滤模型对惯性导航数据进行过滤。
在一些实施例中，不同的过滤模型对应不同的误差阈值范围和参考值；利用过滤模型对惯性导航数据进行过滤包括：根据惯性导航数据和参考值，确定惯性导航数据的误差；根据惯性导航数据的误差与误差阈值范围的比对结果，确定是否对惯性导航数据进行过滤。
在一些实施例中，在参考值为惯性导航数据前一周期惯性导航数据的情况下，利 用过滤模型对惯性导航数据进行过滤包括：根据惯性导航数据与前一周期惯性导航数据的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第一阈值范围的情况下，则将惯性导航数据过滤掉；或者，在参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的情况下，利用过滤模型对惯性导航数据进行过滤包括：从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的差值，并根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第二阈值范围的情况下，将惯性导航数据过滤掉；或者在参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的平均值的情况下，利用过滤模型对惯性导航数据进行过滤包括：根据惯性导航数据与各周期惯性导航数据的平均值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第三阈值范围的情况下，将惯性导航数据过滤掉。
在一些实施例中，在参考值包括惯性导航数据的前一周期惯性导航数据、惯性导航数据的误差率和前一周期惯性导航数据的误差率的情况下，利用过滤模型对惯性导航数据进行过滤包括：根据惯性导航数据的补偿值与前一周期惯性导航数据的补偿值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第四阈值范围的情况下，则将惯性导航数据过滤掉；或者，在参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和各周期的惯性导航数据的误差率的情况下，利用过滤模型对惯性导航数据进行过滤包括：从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的补偿值的差值，根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第五阈值范围的情况下，则将惯性导航数据过滤掉；或者，在参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和各周期的惯性导航数据的误差率的情况下，利用过滤模型对惯性导航数据进行过滤包括：计算惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的补偿值的平均值，根据惯性导航数据的补偿值与平均值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第六阈值范围的情况下，则将惯性导航数据过滤掉；其中，惯性导航数据的补偿值为惯性导航数据与对应的误差率的乘积。
在一些实施例中，该方法还包括：根据被过滤掉的各个惯性导航数据的误差的平均值和对应的参考误差阈值，确定参考误差，利用参考误差修正被过滤掉的各个惯性导航数据；其中，参考误差阈值与过滤模型对应。
在一些实施例中，在根据过滤后的惯性导航数据与激光点云数据构建三维场景之前还包括：将不符合第一预设角度范围的激光束对应的激光点云数据删除；或者将不符合第二预设角度范围的激光束中的激光线对应的激光点云数据删除。
在一些实施例中，根据过滤后的惯性导航数据与激光点云数据构建周围环境三维场景包括：将每一周期获取的激光点云数据由激光坐标系转换至同时刻获取且未被过滤的惯性导航数据的地理位置坐标系；将转换至地理位置坐标系中的各个周期的激光点云数据叠加构建周围环境三维场景。
在一些实施例中，惯性导航数据包括运动姿态数据和地理位置数据中至少一项，运动姿态数据包括航向数据、俯仰数据、横滚数据中至少一项，地理位置数据包括经度、纬度、高程中至少一项；地理位置数据是从卫星导航系统获取并通过惯性导航系统根据惯性方向、运动情况以及各个自由度方向的受力情况进行修正后得到的。
在一些实施例中，测量参数包括：路面状况、载体速度、惯性导航数据的采集频率、惯性导航系统的厂商、惯性导航系统的型号、惯性导航系统硬件误差中的至少一项。
根据本公开的另一些实施例，提供的一种数据处理装置，包括：数据获取单元，被配置为周期性获取载体的惯性导航数据和扫描周围环境得到的激光点云数据；数据过滤单元，被配置为根据测量参数对惯性导航数据进行过滤；建模单元，被配置为根据过滤后的惯性导航数据与激光点云数据构建周围环境的三维场景。
在一些实施例中，数据过滤单元被配置为根据测量参数选取对应的过滤模型，利用过滤模型对惯性导航数据进行过滤。
在一些实施例中，不同的过滤模型对应不同的误差阈值范围和参考值；数据过滤单元被配置为根据惯性导航数据和参考值，确定惯性导航数据的误差，根据惯性导航数据的误差与误差阈值范围的比对结果，确定是否对惯性导航数据进行过滤。
在一些实施例中，在参考值为惯性导航数据的相邻前一周期惯性导航数据的情况下，数据过滤单元被配置为根据惯性导航数据与相邻前一周期惯性导航数据的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第一阈值范围的情况下，则将惯性导航数据过滤掉；或者，在参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的情况下，数据过滤单元被配置为从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的差值，并根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足 第二阈值范围的情况下，将惯性导航数据过滤掉；或者，在参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的平均值的情况下，数据过滤单元被配置为根据惯性导航数据与各周期惯性导航数据的平均值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第三阈值范围的情况下，将惯性导航数据过滤掉。
在一些实施例中，在参考值包括惯性导航数据的前一周期惯性导航数据、惯性导航数据的误差率和前一周期惯性导航数据的误差率的情况下，数据过滤单元被配置为根据惯性导航数据的补偿值与前一周期惯性导航数据的补偿值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第四阈值范围的情况下，则将惯性导航数据过滤掉；或者，在参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和各周期的惯性导航数据的误差率的情况下，数据过滤单元被配置为从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的补偿值的差值，根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第五阈值范围的情况下，则将惯性导航数据过滤掉；或者在参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和各周期的惯性导航数据的误差率的情况下，数据过滤单元被配置为计算惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的补偿值的平均值，根据惯性导航数据的补偿值与平均值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第六阈值范围的情况下，则将惯性导航数据过滤掉；其中，惯性导航数据的补偿值为惯性导航数据与对应的误差率的乘积。
在一些实施例中，该装置还包括：数据修正单元，被配置为根据被过滤掉的各个惯性导航数据的误差的平均值和对应的参考误差阈值，确定参考误差，利用参考误差修正被过滤掉的各个惯性导航数据；其中，参考误差阈值与过滤模型对应。
在一些实施例中，该装置还包括：激光数据筛选单元，被配置为将不符合第一预设角度范围的激光束对应的激光点云数据删除，或者将不符合第二预设角度范围的激光束中的激光线对应的激光点云数据删除。
在一些实施例中，建模单元被配置为将每一周期获取的激光点云数据由激光坐标系转换至同时刻获取且未被过滤的惯性导航数据的地理位置坐标系，将转换为地理位置坐标系的各个周期的激光点云数据叠加构建周围环境三维场景。
在一些实施例中，惯性导航数据包括运动姿态数据和地理位置数据中至少一项， 运动姿态数据包括航向数据、俯仰数据、横滚数据中至少一项，地理位置数据包括经度、纬度、高程中至少一项；地理位置数据是从卫星导航系统获取并通过惯性导航系统根据惯性方向、运动情况以及各个自由度方向的受力情况进行修正后得到的。
在一些实施例中，测量参数包括：路面状况、载体速度、惯性导航数据的采集频率、惯性导航系统的厂商、惯性导航系统的型号、惯性导航系统硬件误差中的至少一项。
根据本公开的又一些实施例，提供的一种数据处理装置，包括：存储器；以及耦接至存储器的处理器，处理器被配置为基于存储在存储器设备中的指令，执行如前述任意实施例的数据处理方法。
根据本公开的再一些实施例，提供的一种计算机可读存储介质，其上存储有计算机程序，其中，该程序被处理器执行时实现前述任意实施例的数据处理方法。
根据本公开的又一些实施例，提供的一种数据处理系统，包括：前述任意实施例的数据处理装置；以及惯性导航系统，被配置为采集惯性导航数据并发送至数据处理装置；激光扫描仪，被配置为采集激光点云数据并发送至数据处理装置。
在一些实施例中，该系统还包括：卫星导航系统，被配置为采集地理位置数据并发送至惯性导航系统；惯性导航系统，还被配置为根据惯性方向、运动情况以及各个自由度方向的受力情况对地理位置数据进行修正。
本公开中根据不同的测量参数，惯性导航数据被过滤，不符合对应的测量参数规律的惯性导航数据能够被过滤。本公开使得在各种测量参数下激光点云数据与惯性导航数据配合构建出的三维场景更加准确。
通过以下参照附图对本公开的示例性实施例的详细描述，本公开的其它特征及其优点将会变得清楚。
此处所说明的附图用来提供对本公开的进一步理解，构成本申请的一部分，本公开的示意性实施例及其说明被配置为解释本公开，并不构成对本公开的不当限定。在附图中：
图1示出本公开的一些实施例的数据处理方法的流程示意图。
图2示出本公开的另一些实施例的数据处理方法的流程示意图。
图3示出本公开的又一些实施例的数据处理方法的流程示意图。
图4示出本公开的再一些实施例的数据处理方法的流程示意图。
图5示出本公开的一些实施例的数据处理装置的结构示意图。
图6示出本公开的另一些实施例的数据处理装置的结构示意图。
图7示出本公开的又一些实施例的数据处理装置的结构示意图。
图8示出本公开的再一些实施例的数据处理装置的结构示意图。
图9示出本公开的又一些实施例的数据处理系统的结构示意图。
下面将结合本公开实施例中的附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本公开一部分实施例，而不是全部的实施例。以下对至少一个示例性实施例的描述实际上仅仅是说明性的，决不作为对本公开及其应用或使用的任何限制。基于本公开中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本公开保护的范围。
针对现有技术中惯性导航数据测量不准确，进而使得构建的周围环境三维场景出现偏差的问题，提出本方案。
下面结合图1描述本公开的数据处理方法的一些实施例。
图1为本公开数据处理方法一些实施例的流程图。如图1所示，该实施例的方法包括：步骤S102～S106。
在步骤S102中，周期性获取载体的惯性导航数据和扫描周围环境得到的激光点云数据。
惯性导航系统和激光扫描扫描仪通常设置于同一载体(例如车辆)之上，随着载体的移动，惯性导航系统和激光扫描扫描仪分别采集载体惯性导航数据和周围环境的激光点云数据，后续将两种数据结合则可以构建周围环境的三维场景。惯性导航数据的采集频率例如为100Hz。惯性导航数据包括运动姿态数据和地理位置数据中至少一项，运动姿态数据包括航向数据、俯仰数据、横滚数据中至少一项，地理位置数据包括经度、纬度、高程中至少一项。
地理位置数据可以是由卫星导航系统，例如GPS(Global Positioning System，全球定位系统)，采集并经过惯性导航系统修正后的地理位置数据。惯性导航系统根据惯性方向、运动情况以及各个自由度方向的受力情况对卫星导航系统采集的地理位置数据进行修正。
当环境对卫星接收产生不良影响，例如树木或楼宇较多，惯性导航系统内置陀螺仪，可以测定运动的惯性方向，进而推算下一时刻自己按照这种惯性将继续运动到达的位置，惯性导航系统负责解算出到达位置的估计坐标。如果没有惯性导航系统，可以在单GPS在信号不好时的情况下，进行辅助定位。
进一步，惯性导航系统也可以有内置加速度计，可以计算空间多自由度方向上受力情况，根据受力导致的姿态改变，及受力大小，对陀螺仪得到的原惯性位置推算结构进行修正。
在载体逐步行进到GPS信号良好区域的情况下，惯性导航系统会将GPS位置替换上述推算的位置。此时由于接收到了较为准确的GPS信号，根据陀螺仪对航位推算误差将被归零。上述过程相对于使用单GPS的定位精度提升很多。
激光扫描仪发出多束激光，多束激光的发射角度不同，每一束激光由多条激光线组成，例如32线，每一线在同一垂直平面内对应不同的发射角度。
在步骤S104中，根据测量参数对惯性导航数据进行过滤。
测量参数例如包括：路面状况、载体速度、惯性导航数据的采集频率、惯性导航系统的厂商、惯性导航系统的型号、惯性导航系统硬件误差中的至少一项。不同的路面状况例如颠簸会造成载体的惯性发生变化，从而导致惯性导航数据的采集出现偏差。惯性导航设备的不同也会导致惯性导航数据的采集出现不同的偏差。针对不同的测量参数，惯性导航数据将呈现不同的变化规律，因此，可以根据实际情况对不同的测量参数建立不同的过滤模型，根据实际的测量参数选取过滤模型对惯性导航数据进行过滤。
在一些实施例中，不同的过滤模型可以对应不同的误差阈值范围和参考值。如图2所示，步骤S104包括：步骤S202～S204。
在步骤S202中，根据惯性导航数据和参考值，确定惯性导航数据的误差。
后续将描述步骤S202中根据惯性导航数据和参考值，确定惯性导航数据的误差的一些实施例。
在步骤S204中，根据惯性导航数据的误差与误差阈值范围的比对结果，确定是否对惯性导航数据进行过滤。
后续将描述利用过滤模型对惯性导航数据进行过滤的一些实施例。
在步骤S106中，根据过滤后的惯性导航数据与激光点云数据构建三维场景。
在一些实施例中，如图3所示，步骤S106包括：步骤S302～步骤S304。
在步骤S302中，将每一周期获取的激光点云数据由激光坐标系转至为同时刻获取的未被过滤的惯性导航数据地理位置坐标系。
在步骤S304中，将转换为地理位置坐标系的各个周期的激光点云数据叠加构建周围环境的三维场景。
激光点云是由激光扫描仪采集得到，原始的三维扫描数据，都是以扫描仪为坐标原点的独立激光坐标系数据。当扫描仪在现实世界产生了运动时，此时多个周期的激光点云数据叠加，会由于应当对应于同一原点，但实际移动后不在同一原点，导致数据叠加后根本无法辨识完整的结果。因此，将扫描仪刚性连接在惯性导航系统上，当扫描仪运动时，惯性导航系统可以获知运动的姿态及位置，将每一周期激光点云数据，对应到同一时刻的运动姿态及位置上，通过将激光点云数据由激光坐标系变换到GPS坐标系，这样可以将多周期的激光点云数据叠加，得到立体完整的三维激光点云数据。激光点云数据处理过程可以经过几个坐标系的转换，首先将激光点云数据从激光坐标系转到笛卡尔坐标系，然后再转到载体坐标系，最终在载体坐标系转成GPS坐标系。
在将不稳定的惯性导航数据过滤掉的同时，可以剔除与这些惯性导航数据同时采集的激光点云数据，使这些激光点云数据不参与最终建模。
上述实施例的方法根据不同的测量参数，惯性导航数据被过滤，不符合对应的测量参数规律的惯性导航数据能够被过滤。上述实施例的方法使得在各种测量参数下激光点云数据与惯性导航数据配合构建出的三维场景更加准确。
上述实施例中，不同的过滤模型对应不同的测量参数，反映不同的惯性导航数据的变化规律，可以根据过滤模型对惯性导航数据进行过滤。本公开提供一些利用过滤模型对惯性导航数据进行过滤的一些实施例。
在一些实施例中，参考值为惯性导航数据的前一周期惯性导航数据。选取的过滤模型对应的误差阈值范围例如为第一阈值范围，第一阈值范围可以表示相邻两次采集的惯性导航数据之间的正常差值范围。第一阈值范围对应于不同类型的惯性导航数据可以包括一个或多个不同的阈值范围，例如包括航向阈值范围、俯仰阈值范围、横滚阈值范围，经度阈值范围、纬度阈值范围、高程阈值范围中至少一项。
进一步，根据惯性导航数据与前一周期惯性导航数据的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第一阈值范围的情况下，则将惯性导航数据过滤掉。例如，计算惯性导航数据与前一次惯性导航数据的差值，作为惯性导航数据的误差，如果惯性导航数据的误差不满足第一阈值范围，则将惯性导航数据过滤掉。
例如，针对某种测量参数，过滤模型对应的相邻两次采集的航向的误差阈值范围为0.01度～0.05度，则分别计算每相邻两次测得的航向的差值，将相邻两次的惯性导航数据中后一次不满足误差阈值范围的剔除。可以从第二次测得的惯性导航数据依次向后对各次惯性导航数据进行过滤，也可以在采集惯性导航数据的同时进行过滤。对于第一次测得的惯性导航数据可以与初始阈值进行比对确定是否进行过滤，该初始阈值与过滤模型相对应，即不同的测量参数下初始阈值可以不同。被过滤掉的惯性导航数据可以不再参与计算，即参考值为未被过滤掉的前一周期惯性导航数据，如果惯性导航数据相邻前N个连续周期的数据全部被过滤掉，N大于预设值，则可以发出警报，停止测量。
在另一些实施例中，参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据，选取的过滤模型对应于第二阈值范围，第二阈值范围例如表示每相邻两次采集的惯性导航数据的平均差值的正常范围。第二阈值范围对应于不同类型的惯性导航数据可以包括一个或多个不同类型的阈值范围，例如包括航向阈值范围、俯仰阈值范围、横滚阈值范围，经度阈值范围、纬度阈值范围、高程阈值范围中至少一项。
进一步，从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两次惯性导航数据的差值，并根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第二阈值范围的情况下，将惯性导航数据过滤掉。可以对各个差值求平均值作为惯性导航数据的误差。
例如，针对第n+1次测得的惯性导航数据value
n+1，其中，n为大于或等于1的正整数，value
n+1例如可以表示俯仰数据，value
n+1的误差可以表示为
i为正整数，假设value
n+1之前有n个未被过滤掉的惯性导航数据。如果X
n+1不满足第二阈值范围，则将value
n+1过滤掉。
在又一些实施例中，参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的平均值。选取的过滤模型对应的误差阈值范围例如为第三阈值范围，第三阈值范围可以表示惯性导航数据与之前的惯性导航数据平均值的正常差值范围。第三阈值范围对应于不同类型的惯性导航数据可以包括一个或多个不同的阈值范围，例如包括航向阈值范围、俯仰阈值范围、横滚阈值范围，经度阈值范围、纬度阈值范围、高程阈值范围中至少一项。
进一步，根据惯性导航数据与各周期惯性导航数据的平均值的差值，确定惯性导 航数据的误差，在惯性导航数据的误差不满足第三阈值范围的情况下，将惯性导航数据过滤掉。例如，将惯性导航数据与各周期惯性导航数据的平均值的差值，确定为惯性导航数据的误差。
参考值也可以为惯性导航数据至第M周期惯性导航数据中未被过滤掉的各周期惯性导航数据的平均值，M为预设值。
在又一些实施例中，参考值包括惯性导航数据前一周期的惯性导航数据、惯性导航数据的误差率和前一周期惯性导航数据的误差率。选取的过滤模型对应的误差阈值范围例如为第四阈值范围，第四阈值范围可以表示相邻两次采集的惯性导航数据的补偿值之间的正常差值范围。由于环境或设备本身造成的误差可以进行一定的补偿，补偿后的惯性导航数据的误差如果还不能满足误差阈值范围，则被过滤掉。惯性导航数据补偿值可以为惯性导航数据与对应的误差率的乘积，误差率根据过滤模型的不同，以及测量周期的不同而不同。第四阈值范围对应于不同类型的惯性导航数据可以包括一个或多个不同的阈值范围，例如包括航向阈值范围、俯仰阈值范围、横滚阈值范围，经度阈值范围、纬度阈值范围、高程阈值范围中至少一项。
进一步，根据惯性导航数据的补偿值与前一周期惯性导航数据的补偿值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第四阈值范围的情况下，则将惯性导航数据过滤掉。例如，惯性导航数据与相邻前一次测得的惯性导航数据分别乘以对应的误差率之后求差值，作为该惯性导航数据的误差。
例如，针对第n+1次测得的惯性导航数据value
n+1，其中，n为大于或等于1的正整数，value
n+1例如可以表示俯仰数据，value
n+1的误差可以表示为X
n+1＝value
n+1·p
n+1-value
n·p
n，其中，p
n+1、p
n分别表示第n+1次和第n次测得的惯性导航数据的误差率，如果X
n+1不满足第四阈值范围，则将value
n+1过滤掉。被过滤掉的惯性导航数据可以不再参与计算。
在另一些实施例中，参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据、各周期的惯性导航数据的误差率。选取的过滤模型对应的误差阈值范围例如为第五阈值范围，第五阈值范围可以表示相邻两次采集的惯性导航数据的补偿值之间的正常平均差值范围。惯性导航数据补偿值可以为惯性导航数据与对应的误差率的乘积，误差率根据过滤模型的不同，以及测量周期的不同而不同。第五阈值范围对应于不同类型的惯性导航数据可以包括一个或多个不同的阈值范围，例如包括航向阈值范围、俯仰阈值范围、横滚阈值范围，经度阈值范围、纬度阈值范 围、高程阈值范围中至少一项。
进一步，从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的补偿值的差值，根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第五阈值范围的情况下，则将惯性导航数据过滤掉。例如，将各个差值的平均值，确定为惯性导航数据的误差。
例如，针对第n+1次测得的惯性导航数据value
n+1，其中，n为大于或等于1的正整数，value
n+1例如可以表示俯仰数据，value
n+1的误差可以表示为
其中，p
i、p
i-1分别表示第i次和第i-1次测得的惯性导航数据的误差率，假设value
n+1之前有n个未被过滤掉的惯性导航数据。如果X
n+1不满足第五阈值范围，则将value
n+1过滤掉。
还可以对相邻两周期的惯性导航数据的差值取绝对值之后再计算各个差值绝对值的均值，作为惯性导航数据的误差。例如value
n+1的误差可以表示为
在又一些实施例中，参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据、各周期的惯性导航数据的误差率。选取的过滤模型对应的误差阈值范围例如为第六阈值范围，第六阈值范围可以表示惯性导航数据补偿值与之前的惯性导航数据的补偿值的平均值的正常差值范围。惯性导航数据补偿值可以为惯性导航数据与对应的误差率的乘积，误差率根据过滤模型的不同，以及测量周期的不同而不同。第六阈值范围对应于不同类型的惯性导航数据可以包括一个或多个不同的阈值范围，例如包括航向阈值范围、俯仰阈值范围、横滚阈值范围，经度阈值范围、纬度阈值范围、高程阈值范围中至少一项。
进一步，计算惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的补偿值的平均值，根据惯性导航数据的补偿值与各周期惯性导航数据的补偿值的平均值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第六阈值范围的情况下，则将惯性导航数据过滤掉。
例如，针对第n+1次测得的惯性导航数据value
n+1，其中，n为大于或等于1的正整数，value
n+1例如可以表示俯仰数据，value
n+1的误差可以表示为
其中，p
i、p
i-1分别表示第i次和第i-1次测得的惯性导航数据的误差率，假设value
n+1之前有n个未被过滤掉的惯性导航数据。如果X
n+1不满足第六阈值范围，则将value
n+1过滤掉。
参考值也可以为惯性导航数据至第M周期惯性导航数据中未被过滤掉的各周期惯性导航数据补偿值的平均值，M为预设值。
在一些实施例，不同的过滤模型对应于不同的参考误差阈值组，参考误差阈值表示过滤掉的各个惯性导航数据的误差的平均值的调整阈值，即过滤掉的各个惯性导航数据的误差的可接受范围。
参考误差阈值组对应于不同的惯性导航数据可以包括一个或多个不同的参考误差阈值，例如包括航向参考误差阈值、俯仰参考误差阈值、横滚参考误差阈值，经度参考误差阈值、纬度参考误差阈值、高程参考误差阈值中至少一项。
根据过滤掉的各个惯性导航数据的误差的平均值和对应的参考误差阈值，确定参考误差，例如计算过滤掉的各个惯性导航数据的误差的平均值，并计算平均值与对应参考误差的差值作为参考误差；利用参考误差修正过滤掉的各个惯性导航数据，即过滤掉的各个惯性导航数据分别减去参考误差。
例如，过滤掉的各个惯性导航数据的误差的平均值与参考误差阈值的差值为e，第n+1次测得的惯性导航数据value
n+1被过滤掉，则value
n+1的修正值为value
n+1-e。被修正的惯性导航数据可以重新被配置为建模数据，弥补数据不足的问题。
上述几种实施例可以任意组合使用，例如，将第一种实施例作为对惯性导航数据的粗略过滤进而用其他实施例进行进一步过滤等。
下面结合图4描述本公开数据处理方法的另一些实施例。
图4为本公开数据处理方法另一些实施例的流程图。如图4所示，在步骤306之前还可以包括：步骤S402～S408。
在步骤S402中，周期性获取惯性导航数据和激光点云数据。
在步骤S404中，根据测量参数对惯性导航数据进行过滤
在步骤S406中，根据测试需求角度对激光束或激光线对应的激光点云数据删除。
在一些实施例中，将不符合第一预设角度范围的激光束对应的激光点云数据删除；或者，将不符合第二预设角度范围的激光束中的激光线对应的激光点云数据删除。
激光扫描仪可以发出平行于地面的360度角度范围的多束激光，每一束激光又可以在垂直于地面的平面内分为多条激光线。因此，可以根据实际测试需求，剔除部分激光束或激光线。例如，只需要测量地面范围内的物体，可以将朝向斜上方的激光线对应的点云数据剔除。只需要测量车辆前进方向左侧的景物，则可以将车辆前进方向右侧的激光束剔除。这样可以节省数据的存储量和运算量，提高效率。
步骤S406可以与步骤S404并列执行没有先后顺序。
在步骤S408中，根据过滤后的惯性导航数据与激光点云数据构建三维场景。
步骤S402、404、408，可以分别参考前述实施例中步骤S102、104、108对应的实施例。
本公开还提供一种数据处理装置，下面结合图5进行描述。
图5为本公开数据处理装置一些实施例的结构图。如图5所示，该装置50包括：数据获取单元502，数据过滤单元504，建模单元506。
数据获取单元502，被配置为周期性获取载体的惯性导航数据和扫描周围环境得到的激光点云数据。数据获取单元502可以执行步骤S102的方法。
惯性导航数据包括运动姿态数据和地理位置数据中至少一项，运动姿态数据包括航向数据、俯仰数据、横滚数据中至少一项，地理位置数据包括经度、纬度、高程中至少一项；地理位置数据是从卫星导航系统获取并通过惯性导航系统根据惯性方向、运动情况以及各个自由度方向的受力情况进行修正后得到的。
数据过滤单元504，被配置为根据测量参数对惯性导航数据进行过滤。数据过滤单元504可以执行步骤S104的方法。
测量参数例如包括：路面状况、载体速度、惯性导航数据的采集频率、惯性导航系统的厂商、惯性导航系统的型号、惯性导航系统硬件误差中的至少一项。
在一些实施例中，数据过滤单元504被配置为根据测量参数选取对应的过滤模型，利用过滤模型对惯性导航数据进行过滤。
在一些实施例中，不同的过滤模型对应不同的误差阈值范围和参考值。数据过滤单元504被配置为根据惯性导航数据和参考值，确定惯性导航数据的误差，根据惯性导航数据的误差与误差阈值范围的比对结果，确定是否对惯性导航数据进行过滤。
在一些实施例中，在参考值为惯性导航数据相邻前一周期的惯性导航数据的情况下，数据过滤单元504被配置为根据惯性导航数据与相邻前一周期惯性导航数据的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第一阈值范围的情况下，则将惯性导航数据过滤掉。
在一些实施例中，在参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的情况下，数据过滤单元504被配置为从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两次惯性导航数据的差值，并根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第二 阈值范围的情况下，将惯性导航数据过滤掉。
在一些实施例中，在参考值为惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的平均值的情况下，数据过滤单元504被配置为根据惯性导航数据与各周期惯性导航数据的平均值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第三阈值范围的情况下，将惯性导航数据过滤掉。
在一些实施例中，在参考值包括惯性导航数据相邻前一周期惯性导航数据、惯性导航数据的误差率和相邻前一周期惯性导航数据的误差率的情况下，数据过滤单元504被配置为根据惯性导航数据的补偿值与相邻前一周期惯性导航数据的补偿值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第四阈值范围的情况下，则将惯性导航数据过滤掉。
在一些实施例中，在参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据、各周期的惯性导航数据的误差率的情况下，数据过滤单元504被配置为从惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的补偿值的差值，根据各个差值的平均值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第五阈值范围的情况下，则将惯性导航数据过滤掉。
在一些实施例中，在参考值包括惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据、各周期的惯性导航数据的误差率的情况下，数据过滤单元504被配置为计算惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的补偿值的平均值，根据惯性导航数据的补偿值与各周期惯性导航数据的补偿值的平均值的差值，确定惯性导航数据的误差，在惯性导航数据的误差不满足第六阈值范围的情况下，则将惯性导航数据过滤掉。惯性导航数据的补偿值为惯性导航数据与对应的误差率的乘积。
建模单元506，被配置为根据过滤后的惯性导航数据与激光点云数据构建三维场景。建模单元506可以执行步骤S106的方法。
在一些实施例中，建模单元506，被配置为将每一周期获取的激光点云数据由激光坐标系转换至同时刻获取且未被过滤的惯性导航数据的地理位置坐标系，将转换为地理位置坐标系的各个周期的激光点云数据叠加构建周围环境三维场景。
下面结合图6进行描述数据处理装置的另一些实施例。
图6为本公开数据处理装置一些实施例的结构图。如图6所示，该装置60包括： 数据获取单元602，数据过滤单元604，建模单元606，这三个单元分别可以实现图5中数据获取单元502，数据过滤单元504，建模单元506类似的功能。
数据修正单元608，被配置根据过滤掉的各个惯性导航数据的误差的平均值和对应的参考误差阈值，确定参考误差，利用参考误差修正过滤掉的各个惯性导航数据。参考误差阈值与过滤模型对应。
进一步，数据处理装置60还可以包括：激光数据筛选单元610，被配置为将不符合第一预设角度范围的激光束对应的激光点云数据删除，或者将不符合第一预设角度范围的激光束对应的激光点云数据删除，或者将不符合第二预设角度范围的激光束中的激光线对应的激光点云数据删除。
本公开的实施例中的数据处理装置可各由各种计算设备或计算机系统来实现，下面结合图7以及图8进行描述。
图7为本公开数据处理装置的一些实施例的结构图。如图7所示，该实施例的装置70包括：存储器710以及耦接至该存储器710的处理器720，处理器720被配置为基于存储在存储器710中的指令，执行本公开中任意一些实施例中的数据处理方法。
其中，存储器710例如可以包括系统存储器、固定非易失性存储介质等。系统存储器例如存储有操作系统、应用程序、引导装载程序(Boot Loader)、数据库以及其他程序等。
图8为本公开数据处理装置的另一些实施例的结构图。如图2所示，该实施例的装置80包括：存储器810以及处理器820，分别与图7中存储器710以及处理器720类似。还可以包括输入输出接口830、网络接口840、存储接口850等。这些接口830，840，850以及存储器810和处理器820之间例如可以通过总线860连接。其中，输入输出接口830为显示器、鼠标、键盘、触摸屏等输入输出设备提供连接接口。网络接口840为各种联网设备提供连接接口，例如可以连接到数据库服务器或者云端存储服务器等。存储接口850为SD卡、U盘等外置存储设备提供连接接口。
本公开还提供一种数据处理系统，下面结合图9进行描述。
图9为本公开数据处理系统一些实施例的结构图。如图9所示，该系统9包括：前述任一些实施例的数据处理装置50、60、70或80，以及惯性导航系统92，激光扫描仪94。
惯性导航系统92，被配置为采集惯性导航数据并发送至数据处理装置。
激光扫描仪94，被配置为采集激光点云数据并发送至数据处理装置。
在一些实施例中，该系统9还可以包括：卫星导航系统96，被配置为采集地理位置数据并发送至惯性导航系统92。
惯性导航系统92，还被配置为根据惯性方向、运动情况以及各个自由度方向的受力情况对地理位置数据进行修正。
本公开还提供一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现前述任一些实施例的数据处理方法的步骤。
本领域技术人员可以理解，本公开的数据处理方法、装置、系统除了被配置为进行测绘工作外，可以应被配置为无人驾驶领域，无人驾驶车可以根据本公开的数据处理方法、装置、系统构建的三维地图，从而识别驾驶环境实现无人驾驶。此外，本公开的数据处理方法、装置、系统还可以应被配置为无人仓储领域，无人仓储中搬运车可以通过构建的三维仓库地图，完成货物的搬运等。
本领域内的技术人员应当明白，本公开的实施例可提供为方法、系统、或计算机程序产品。因此，本公开可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本公开可采用在一个或多个其中包含有计算机可用程序代码的计算机可用非瞬时性存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本公开是参照根据本公开实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解为可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生被配置为实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供被配置为实现在流程图一个流程或多个流程 和/或方框图一个方框或多个方框中指定的功能的步骤。
以上所述仅为本公开的较佳实施例，并不用以限制本公开，凡在本公开的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本公开的保护范围之内。
Claims (24)
- 一种数据处理方法，包括：周期性获取载体的惯性导航数据和扫描周围环境得到的激光点云数据；根据测量参数对所述惯性导航数据进行过滤；根据过滤后的惯性导航数据与所述激光点云数据构建周围环境的三维场景。
- 根据权利要求1所述的数据处理方法，其中，所述根据测量参数对所述惯性导航数据进行过滤包括：根据测量参数选取对应的过滤模型；利用所述过滤模型对所述惯性导航数据进行过滤。
- 根据权利要求2所述的数据处理方法，其中，不同的过滤模型对应不同的误差阈值范围和参考值；所述利用所述过滤模型对所述惯性导航数据进行过滤包括：根据惯性导航数据和参考值，确定所述惯性导航数据的误差；根据所述惯性导航数据的误差与误差阈值范围的比对结果，确定是否对所述惯性导航数据进行过滤。
- 根据权利要求3所述的数据处理方法，其中，在所述参考值为所述惯性导航数据前一周期惯性导航数据的情况下，所述利用所述过滤模型对所述惯性导航数据进行过滤包括：根据所述惯性导航数据与前一周期惯性导航数据的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第一阈值范围的情况下，则将所述惯性导航数据过滤掉；或者在所述参考值为所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的情况下，所述利用所述过滤模型对所述惯性导航数据进行过滤包括：从所述惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的差值，并根据各个差值的平均值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第二阈值范围的情况下，将所述惯性导航数据过滤掉；或者在所述参考值为所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各 周期惯性导航数据的平均值的情况下，所述利用所述过滤模型对所述惯性导航数据进行过滤包括：根据所述惯性导航数据与所述各周期惯性导航数据的平均值的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第三阈值范围的情况下，将所述惯性导航数据过滤掉。
- 根据权利要求3所述的数据处理方法，其中，在所述参考值包括所述惯性导航数据前一周期惯性导航数据、所述惯性导航数据的误差率和所述前一周期惯性导航数据的误差率的情况下，所述利用所述过滤模型对所述惯性导航数据进行过滤包括：根据所述惯性导航数据的补偿值与前一周期惯性导航数据的补偿值的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第四阈值范围的情况下，则将所述惯性导航数据过滤掉；或者在所述参考值包括所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和所述各周期的惯性导航数据的误差率的情况下，所述利用所述过滤模型对所述惯性导航数据进行过滤包括：从所述惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的补偿值的差值，根据各个差值的平均值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第五阈值范围的情况下，则将所述惯性导航数据过滤掉；或者在所述参考值包括所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和所述各周期的惯性导航数据的误差率的情况下，所述利用所述过滤模型对所述惯性导航数据进行过滤包括：计算所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的补偿值的平均值，根据所述惯性导航数据的补偿值与所述平均值的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第六阈值范围的情况下，则将所述惯性导航数据过滤掉；其中，惯性导航数据的补偿值为惯性导航数据与对应的误差率的乘积。
- 根据权利要求3所述的数据处理方法，还包括：根据被过滤掉的各个惯性导航数据的误差的平均值和对应的参考误差阈值，确定参考误差；利用所述参考误差修正所述被过滤掉的各个惯性导航数据；其中，所述参考误差阈值与所述过滤模型对应。
- 根据权利要求1所述的数据处理方法，其中，在根据过滤后的惯性导航数据与所述激光点云数据构建三维场景之前还包括：将不符合第一预设角度范围的激光束对应的激光点云数据删除；或者将不符合第二预设角度范围的激光束中的激光线对应的激光点云数据删除。
- 根据权利要求1所述的数据处理方法，其中，所述根据过滤后的惯性导航数据与所述激光点云数据构建周围环境三维场景包括：将每一周期获取的激光点云数据由激光坐标系转换至同时刻获取且未被过滤的惯性导航数据的地理位置坐标系；将转换至地理位置坐标系中的各个周期的激光点云数据叠加构建周围环境三维场景。
- 根据权利要求1所述的数据处理方法，其中，所述惯性导航数据包括运动姿态数据和地理位置数据中至少一项，所述运动姿态数据包括航向数据、俯仰数据、横滚数据中至少一项，所述地理位置数据包括经度、纬度、高程中至少一项；所述地理位置数据是从卫星导航系统获取并通过惯性导航系统根据惯性方向、运动情况以及各个自由度方向的受力情况进行修正后得到的。
- 根据权利要求1-9任一项所述的数据处理方法，其中，所述测量参数包括：路面状况、载体速度、惯性导航数据的采集频率、惯性导航系统的厂商、惯性导航系统的型号、惯性导航系统硬件误差中的至少一项。
- 一种数据处理装置，包括：数据获取单元，被配置为周期性获取载体的惯性导航数据和扫描周围环境得到的激光点云数据；数据过滤单元，被配置为根据测量参数对所述惯性导航数据进行过滤；建模单元，被配置为根据过滤后的惯性导航数据与所述激光点云数据构建周围环境的三维场景。
- 根据权利要求11所述的数据处理装置，其中，所述数据过滤单元被配置为根据测量参数选取对应的过滤模型，利用所述过滤模 型对所述惯性导航数据进行过滤。
- 根据权利要求12所述的数据处理装置，其中，不同的过滤模型对应不同的误差阈值范围和参考值；所述数据过滤单元被配置为根据惯性导航数据和参考值，确定所述惯性导航数据的误差，根据所述惯性导航数据的误差与误差阈值范围的比对结果，确定是否对所述惯性导航数据进行过滤。
- 根据权利要求13所述的数据处理装置，其中，在所述参考值为所述惯性导航数据相邻前一周期惯性导航数据的情况下，所述数据过滤单元被配置为根据所述惯性导航数据与相邻前一周期惯性导航数据的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第一阈值范围的情况下，则将所述惯性导航数据过滤掉；或者在所述参考值为所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的情况下，所述数据过滤单元被配置为从所述惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的差值，并根据各个差值的平均值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第二阈值范围的情况下，将所述惯性导航数据过滤掉；或者在所述参考值为所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的平均值的情况下，所述数据过滤单元被配置为根据所述惯性导航数据与所述各周期惯性导航数据的平均值的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第三阈值范围的情况下，将所述惯性导航数据过滤掉。
- 根据权利要求13所述的数据处理装置，其中，在所述参考值包括所述惯性导航数据相邻前一周期惯性导航数据、所述惯性导航数据的误差率和所述前一周期惯性导航数据的误差率的情况下，所述数据过滤单元被配置为根据所述惯性导航数据的补偿值与相邻前一周期惯性导航数据的补偿值的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第四阈值范围的情况下，则将所述惯性导航数据过滤掉；或者，在所述参考值包括所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和所述各周期的惯性导航数据的误差率的情况下，所述数据过滤单元被配置为从所述惯性导航数据至第一周期惯性导航数据，分别计算未被过滤掉的每相邻两周期惯性导航数据的补偿值的差值，根据各个差值的平均值，确定所述惯性 导航数据的误差，在所述惯性导航数据的误差不满足第五阈值范围的情况下，则将所述惯性导航数据过滤掉；或者在所述参考值包括所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据和所述各周期的惯性导航数据的误差率的情况下，所述数据过滤单元被配置为计算所述惯性导航数据至第一周期惯性导航数据中未被过滤掉的各周期惯性导航数据的补偿值的平均值，根据所述惯性导航数据的补偿值与所述平均值的差值，确定所述惯性导航数据的误差，在所述惯性导航数据的误差不满足第六阈值范围的情况下，则将所述惯性导航数据过滤掉；其中，惯性导航数据的补偿值为惯性导航数据与对应的误差率的乘积。
- 根据权利要求13所述的数据处理装置，还包括：数据修正单元，被配置为根据被过滤掉的各个惯性导航数据的误差的平均值和对应的参考误差阈值，确定参考误差，利用所述参考误差修正所述被过滤掉的各个惯性导航数据；其中，所述参考误差阈值与所述过滤模型对应。
- 根据权利要求11所述的数据处理装置，还包括：激光数据筛选单元，被配置为将不符合第一预设角度范围的激光束对应的激光点云数据删除，或者将不符合第二预设角度范围的激光束中的激光线对应的激光点云数据删除。
- 根据权利要求11所述的数据处理装置，其中，所述建模单元被配置为将每一周期获取的激光点云数据由激光坐标系转换至同时刻获取且未被过滤的惯性导航数据的地理位置坐标系，将转换至地理位置中坐标系的各个周期的激光点云数据叠加构建周围环境三维场景。
- 根据权利要求11所述的数据处理装置，其中，所述惯性导航数据包括运动姿态数据和地理位置数据中至少一项，所述运动姿态数据包括航向数据、俯仰数据、横滚数据中至少一项，所述地理位置数据包括经度、纬度、高程中至少一项；所述地理位置数据是从卫星导航系统获取并通过惯性导航系统根据惯性方向、运动情况以及各个自由度方向的受力情况进行修正后得到的。
- 根据权利要求11-19任一项所述的数据处理装置，其中，所述测量参数包括：路面状况、载体速度、惯性导航数据的采集频率、惯性导航 系统的厂商、惯性导航系统的型号、惯性导航系统硬件误差中的至少一项。
- 一种数据处理装置，包括：存储器；以及耦接至所述存储器的处理器，所述处理器被配置为基于存储在所述存储器设备中的指令，执行如权利要求1-10任一项所述的数据处理方法。
- 一种计算机可读存储介质，其上存储有计算机程序，其中，该程序被处理器执行时实现权利要求1-10任一项所述的数据处理方法的步骤。
- 一种数据处理系统，包括：权利要求11-21任一项所述的数据处理装置；以及惯性导航系统，被配置为采集惯性导航数据并发送至所述数据处理装置；激光扫描仪，被配置为采集激光点云数据并发送至所述数据处理装置。
- 根据权利要求19所述的数据处理系统，还包括：卫星导航系统，被配置为采集地理位置数据并发送至所述惯性导航系统；所述惯性导航系统，还被配置为根据惯性方向、运动情况以及各个自由度方向的受力情况对所述地理位置数据进行修正。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710700605.5A CN107527382B (zh) | 2017-08-16 | 2017-08-16 | 数据处理方法以及装置 |
| CN201710700605.5 | 2017-08-16 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2019033882A1 true WO2019033882A1 (zh) | 2019-02-21 |
Family
ID=60681323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2018/095740 WO2019033882A1 (zh) | 2017-08-16 | 2018-07-16 | 数据处理方法、装置、系统和计算机可读存储介质 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN107527382B (zh) |
| WO (1) | WO2019033882A1 (zh) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107527382B (zh) * | 2017-08-16 | 2020-11-03 | 北京京东尚科信息技术有限公司 | 数据处理方法以及装置 |
| CN108225258A (zh) * | 2018-01-09 | 2018-06-29 | 天津大学 | 基于惯性单元和激光跟踪仪动态位姿测量装置和方法 |
| CN111207740A (zh) * | 2020-01-13 | 2020-05-29 | 北京京东乾石科技有限公司 | 车辆定位的方法、装置、设备和计算机可读介质 |
| CN111898094B (zh) * | 2020-07-06 | 2021-04-06 | 广州市吉华勘测股份有限公司 | 基坑监测数据的处理方法、装置、电子设备及存储介质 |
| CN112648008B (zh) * | 2020-12-28 | 2023-09-29 | 北京宸控科技有限公司 | 一种液压支架搬运方法 |
| CN113945213B (zh) * | 2021-09-22 | 2022-05-27 | 北京连山科技股份有限公司 | 一种基于惯性组合导航数据的预估校正方法 |
| CN114046792B (zh) * | 2022-01-07 | 2022-04-26 | 陕西欧卡电子智能科技有限公司 | 一种无人船水面定位与建图方法、装置及相关组件 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103148803A (zh) * | 2013-02-28 | 2013-06-12 | 中国地质大学（北京） | 轻小型三维激光扫描测量系统及方法 |
| CN105261060A (zh) * | 2015-07-23 | 2016-01-20 | 东华大学 | 基于点云压缩和惯性导航的移动场景实时三维重构方法 |
| US20170016870A1 (en) * | 2012-06-01 | 2017-01-19 | Agerpoint, Inc. | Systems and methods for determining crop yields with high resolution geo-referenced sensors |
| CN106504275A (zh) * | 2016-10-12 | 2017-03-15 | 杭州深瞳科技有限公司 | 一种惯性定位与点云配准耦合互补的实时三维重建方法 |
| CN107527382A (zh) * | 2017-08-16 | 2017-12-29 | 北京京东尚科信息技术有限公司 | 数据处理方法以及装置 |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8380375B2 (en) * | 2009-08-14 | 2013-02-19 | IPOZ Systems, LLC | Device, computer storage medium, and computer implemented method for metrology with inertial navigation system and aiding |
| CN104268933B (zh) * | 2014-09-11 | 2017-02-15 | 大连理工大学 | 一种车载二维激光运动中三维环境扫描成像方法 |
| CN104501829B (zh) * | 2014-11-24 | 2017-04-12 | 杭州申昊科技股份有限公司 | 一种惯性导航系统的误差校正方法 |
-
2017
- 2017-08-16 CN CN201710700605.5A patent/CN107527382B/zh active Active
-
2018
- 2018-07-16 WO PCT/CN2018/095740 patent/WO2019033882A1/zh active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170016870A1 (en) * | 2012-06-01 | 2017-01-19 | Agerpoint, Inc. | Systems and methods for determining crop yields with high resolution geo-referenced sensors |
| CN103148803A (zh) * | 2013-02-28 | 2013-06-12 | 中国地质大学（北京） | 轻小型三维激光扫描测量系统及方法 |
| CN105261060A (zh) * | 2015-07-23 | 2016-01-20 | 东华大学 | 基于点云压缩和惯性导航的移动场景实时三维重构方法 |
| CN106504275A (zh) * | 2016-10-12 | 2017-03-15 | 杭州深瞳科技有限公司 | 一种惯性定位与点云配准耦合互补的实时三维重建方法 |
| CN107527382A (zh) * | 2017-08-16 | 2017-12-29 | 北京京东尚科信息技术有限公司 | 数据处理方法以及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107527382B (zh) | 2020-11-03 |
| CN107527382A (zh) | 2017-12-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2019033882A1 (zh) | 数据处理方法、装置、系统和计算机可读存储介质 | |
| CN111156998B (zh) | 一种基于rgb-d相机与imu信息融合的移动机器人定位方法 | |
| KR102063534B1 (ko) | 라이다를 이용한 지도 생성 방법 | |
| CN109710724B (zh) | 一种构建点云地图的方法和设备 | |
| KR102007567B1 (ko) | 스테레오 드론 및 이를 이용하는 무기준점 토공량 산출 방법과 시스템 | |
| CN109084732A (zh) | 定位与导航方法、装置及处理设备 | |
| US20210183100A1 (en) | Data processing method and apparatus | |
| CN112219087A (zh) | 位姿预测方法、地图构建方法、可移动平台及存储介质 | |
| KR101909953B1 (ko) | 라이다 센서를 이용한 차량의 자세 추정 방법 | |
| KR102130687B1 (ko) | 다중 센서 플랫폼 간 정보 융합을 위한 시스템 | |
| JP7431320B2 (ja) | デジタル地図データを使用する方法およびシステム | |
| CN109146990B (zh) | 一种建筑轮廓的计算方法 | |
| CN111708048A (zh) | 点云的运动补偿方法、装置和系统 | |
| KR20190104304A (ko) | 점군을 포함하는 라이다 지도 생성 방법 | |
| CN112154303A (zh) | 高精度地图定位方法、系统、平台及计算机可读存储介质 | |
| KR102115004B1 (ko) | 항공사진을 이용하여 3차원 지도를 생성하는 장치 및 방법 | |
| CN111862214A (zh) | 计算机设备定位方法、装置、计算机设备和存储介质 | |
| WO2024062781A1 (ja) | 演算装置、演算方法およびプログラム | |
| CN116380039A (zh) | 一种基于固态激光雷达和点云地图的移动机器人导航系统 | |
| WO2024057757A1 (ja) | 演算装置、演算方法およびプログラム | |
| US20210264666A1 (en) | Method for obtaining photogrammetric data using a layered approach | |
| JP2003247805A (ja) | 体積計測方法及び体積計測プログラム | |
| Karam et al. | Integrating a low-cost mems imu into a laser-based slam for indoor mobile mapping | |
| KR102050995B1 (ko) | 공간좌표의 신뢰성 평가 장치 및 방법 | |
| CN108253942B (zh) | 一种提高倾斜摄影测量空三质量的方法 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18846424 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 19/08/2020) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 18846424 Country of ref document: EP Kind code of ref document: A1 |