CN105654516B - Satellite image based on target conspicuousness is to ground weak moving target detection method - Google Patents
Satellite image based on target conspicuousness is to ground weak moving target detection method Download PDFInfo
- Publication number
- CN105654516B CN105654516B CN201610091741.4A CN201610091741A CN105654516B CN 105654516 B CN105654516 B CN 105654516B CN 201610091741 A CN201610091741 A CN 201610091741A CN 105654516 B CN105654516 B CN 105654516B
- Authority
- CN
- China
- Prior art keywords
- track
- image
- target
- moving
- frame
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 29
- 239000000203 mixture Substances 0.000 claims abstract description 5
- 238000001914 filtration Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 230000002401 inhibitory effect Effects 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 17
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000005259 measurement Methods 0.000 abstract 1
- 238000012795 verification Methods 0.000 abstract 1
- 238000004364 calculation method Methods 0.000 description 3
- 235000006629 Prosopis spicigera Nutrition 0.000 description 2
- 240000000037 Prosopis spicigera Species 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003014 reinforcing effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20004—Adaptive image processing
- G06T2207/20012—Locally adaptive
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of satellite images based on target conspicuousness to ground weak moving target detection method, for solving the existing method technical problem low to weak moving target detection rate.Technical solution is to carry out background modeling to Saliency maps picture first, carries out significance analysis to gray level image, strengthens the moving target in image.By carrying out Gaussian Mixture background modeling to Saliency maps picture, solves weak moving target detection problem, generate track using classification, filter out false-alarm using trace information, with respect to the background art method, verification and measurement ratio improves 10%, and false alarm rate reduces 5%.
Description
Technical Field
The invention relates to a method for detecting a satellite image-to-ground weak and small moving target, in particular to a method for detecting a satellite image-to-ground weak and small moving target based on target significance.
Background
Satellite image earth moving object detection is an important research topic in the field of computer vision. The existing satellite image detection algorithm for the earth weak moving target has low detection rate. The document "Vehicle Detection and road surface Removal in High Resolution Satellite Images, 2010" discloses a method for detecting a Satellite image on a ground weak and small moving target. The method detects the elliptical moving vehicle in the satellite image by using the geometric characteristics of the vehicle in the satellite image and adopting an improved scale circular spot matching algorithm. For vehicles running on the road, Gaussian elliptical Laplace filtering is carried out, the direction of the road in the image is consistent with the geometric direction of the elliptical moving target, and when the analytic expression of the filtering equation reaches a local extreme value, a candidate moving target in the image can be obtained. However, the detection result of the method depends on the characteristics of the geometric direction of the moving target, the geometric direction of the moving target cannot be obtained for the weak moving target, the detection rate of the weak moving target is extremely low, and the false alarm rate is high.
Disclosure of Invention
In order to overcome the defect that the existing method has low detection rate on the weak and small moving target, the invention provides a method for detecting the weak and small moving target on the ground based on a satellite image of target significance. The method comprises the steps of firstly carrying out background modeling on a saliency image, carrying out saliency analysis on a gray level image, and reinforcing a moving target in the image. The method solves the problem of weak and small moving target detection by performing Gaussian mixture background modeling on the significant image, generates tracks by classification, and filters out false alarms by utilizing track information, so that the detection rate is improved by 10% and the false alarm rate is reduced by 5% compared with a background technology method.
The technical scheme adopted by the invention for solving the technical problems is as follows: a satellite image ground weak and small moving target detection method based on target saliency is characterized by comprising the following steps:
step one, calculating a saliency image for a currently input image. Recording a certain pixel value in the image as IkWhere k represents the kth pixel in the image. Let w be pixel IkA sliding window in the center. | l | · | | denotes an inter-pixel gray scale distance matrix, diRepresenting the distance of pixel k to pixel i. Then the image pixel IkIs expressed as a local saliency value of
Wherein f (-) is with respect to diA linear decreasing function of.
When the gray difference between a certain pixel and its adjacent pixels is larger, the probability that the pixel is a moving object is higher, and therefore the significance multiple of the pixel is smaller along with the increase of the distance. Obtaining each pixel I by formula (1) in the input gray imagekAnd (4) local area saliency, and finally forming a local saliency image.
And secondly, obtaining a background image of the input image by utilizing self-adaptive Gaussian mixture background modeling on the saliency image, and performing background subtraction according to the background image and the input saliency image to obtain a candidate moving target set.
And step three, filtering the false alarms of the candidate moving target set based on the space-time information, and inhibiting the false alarms by using the target continuity information. The association of the trajectory segments is generated hierarchically based on the motion, trajectory and spatiotemporal information. The track associations are divided into primary and advanced tracklets associations.
Primary track association: suppose that there are m targets in the t-th frame, n targets in the t + 1-th frame, and the targets in the t-th frame and the targets in the t + 1-th frame are in a one-to-one relationship. And taking the Euclidean distance between the target in the current frame and each target in the next frame as an input matrix, and calculating the target association information of the t-th frame and the t + 1-th frame by using a Hungarian algorithm. If m < n, namely more moving objects exist in the t +1 th frame, the moving objects which are not related are used as new moving objects. If m is larger than or equal to n, namely more targets exist in the t-th frame, the targets are reserved for two frames, and then whether the false alarm exists is judged according to subsequent association.
High-level tracklets associations: and forming a short track set by using detection points extracted from each frame by using the Hungarian algorithm. The accuracy of the long trajectory is guaranteed by using motion similarity constraints, trajectory similarity constraints and space-time constraints.
Motion similarity constraint: linear function ofFunction representing the distance of movement of the object, representationTrack segment viThe mth position in the time interval, at represents the time interval, and v represents the velocity constant during this time interval. Order toPresentation detectionAndthe time interval in between. Then temporally adjacent tracks viAnd vjRespectively predicting backwards and forwards, and only calculating tau frame to reduce calculation complexity, and tracing trackAndthe motion similarities between are:
wherein dis1And dis2Respectively as follows:
and (3) constraint of track similarity: the track of a vehicle moving on a road is a straight line in a short time domain. A straight line is thus fitted from the short trajectory using the RANSAC algorithm. Trajectory viShould be distributed over the trajectory viEstimated straight line LiOr in the vicinity of this line. The trace obtained from the false alarm point association is discarded. If two tracks belong to the same target, two tracks viAnd vjThe track similarity between them is expressed as
Wherein,is the midpoint of the trace jAnd a straight line LiThe distance of the straight line.
Space-time constraint: since the vehicle cannot appear at two different locations at the same time, moving objects at different locations in the same frame are necessarily not the same object. A vehicle moving on the road must not be associated with an object in the current field of view if it goes out of range of the satellite photograph, and the trajectory section viAnd vjIf there is an overlapping area or one of the associated tracks already exists, the association between the two tracks is S (v)i,vj)＝0。
From the three constraints, the similarity constraint between two trajectory segments is obtained as follows:
S(vi,vj)＝w1Sm(vi,vj)+w2St(vi,vj) (6)
wherein S ism(. and S)t(. cndot.) represents motion similarity and trajectory similarity, respectively. w is a1And w2Is its weighting factor.
And after the similarity among different tracks is obtained, the candidate target tracks are formed by utilizing the Hungarian algorithm in a correlation mode, and the tracks after the correlation are not straight lines due to the fact that detection points generated by false alarms are ignored, so that straight line fitting is carried out on the generated tracks. Counting the number of points on the straight line according to the best matching degree, namely the matching degree of the inner point rate which is in line with the obtained linear equation in the track and the target track set and the calculated model, and keeping the point with the best matching degreeParameters of the linear equation. When the track length exceeds TlengthAnd the inner point rate is higher than TinnerThen, the moving object on the track is considered as the real object.
Therefore, weak and small moving objects in the satellite images are obtained, the moving objects which are filtered based on the space-time information false alarm are used as real moving objects, and the target results are output in the images.
The invention has the beneficial effects that: the method comprises the steps of firstly carrying out background modeling on a saliency image, carrying out saliency analysis on a gray level image, and reinforcing a moving target in the image. The method solves the problem of weak and small moving target detection by performing Gaussian mixture background modeling on the significant image, generates tracks by classification, and filters out false alarms by utilizing track information, so that the detection rate is improved by 10% and the false alarm rate is reduced by 5% compared with a background technology method.
The present invention will be described in detail with reference to the following embodiments.
Detailed Description
The method for detecting the weak and small moving target on the ground based on the satellite image of the target significance comprises the following specific steps:
(a) and (3) calculating a saliency image: a saliency image is calculated for a currently input image. Recording a certain pixel value in the image as IkWhere k represents the kth pixel in the image. Let w be pixel IkA sliding window in the center. | l | · | | denotes an inter-pixel gray scale distance matrix, diRepresenting the distance of pixel k to pixel i. Then the image pixel IkIs expressed as a local saliency value of
Wherein f (-) is with respect to diA linear decreasing function of.
A segment of image sequence shot by a satellite is input, and the local saliency of each pixel is calculated for a current frame image by using the local saliency. In order to ensure the calculation efficiency and the detection performance, the size of the sliding window w is 3 in this example.
When the gray difference between a certain pixel and its adjacent pixels is larger, the probability that the pixel is a moving object is higher, and therefore the significance multiple of the pixel is smaller along with the increase of the distance. For an input gray image, each pixel I can be obtained by the formula (1)kAnd (4) local area saliency, and finally forming a local saliency image.
(b) Modeling Gaussian mixed background: because the mixed Gaussian background modeling has stronger adaptability to the complex background, the background image of the input image is obtained by utilizing the self-adaptive mixed Gaussian background modeling for the saliency image, and the background subtraction is carried out according to the background image and the input saliency image to obtain the candidate moving target set.
And calculating a background image by using an adaptive mixed Gaussian background modeling method for the input local saliency image in consideration of the calculation complexity and accuracy. The background distribution of each pixel point is approximated in this example by a gaussian distribution of 3 different weights. An initial background model is then computed using the first ten consecutive images. The image obtained by the background difference is thresholded, and the threshold is set to 15. Thus, a candidate moving object set is obtained.
(c) False alarm filtering based on space-time information: and considering that the appearance characteristics of the moving target are unavailable, filtering the false alarm of the candidate moving target set based on the spatio-temporal information, and inhibiting the false alarm by using the target continuity information. The association of the trajectory segments is generated hierarchically based on the motion, trajectory and spatiotemporal information. The track associations are divided into primary and advanced tracklets associations.
Primary track association: suppose that there are m targets in the t-th frame, n targets in the t + 1-th frame, and the targets in the t-th frame and the targets in the t + 1-th frame are in a one-to-one relationship. And taking the Euclidean distance between the target in the current frame and each target in the next frame as an input matrix, and calculating the target association information of the t-th frame and the t + 1-th frame by using a Hungarian algorithm. The number of moving objects may differ between the upper and lower frames due to the presence of false alarms. If m < n, that is, there are more possible moving objects in the t +1 th frame, the moving object not associated is used as a new moving object. If m is larger than or equal to n, namely more targets exist in the t-th frame, the targets are reserved for two frames, and then whether the false alarm exists is judged according to subsequent association.
High-level tracklets associations: and forming a short track set by using detection points extracted from each frame by using the Hungarian algorithm. The accuracy of the long trajectory is guaranteed by using motion similarity constraints, trajectory similarity constraints and space-time constraints.
Motion similarity constraint: linear function ofFunction representing the distance of movement of the object, representationTrack segment viThe mth position in the time interval, at represents the time interval, and v represents the velocity constant during this time interval. Order toPresentation detectionAndthe time interval in between. Then temporally adjacent tracks viAnd vjRespectively predicting backwards and forwards, and in order to reduce the computational complexity, only computing the tau frame and then the trackAndthe motion similarities between are:
wherein dis1And dis2Respectively as follows:
and (3) constraint of track similarity: the track of a vehicle moving on a road is a straight line in a short time domain. A straight line is thus fitted from the short trajectory using the RANSAC algorithm. Trajectory viShould be distributed over the trajectory viEstimated straight line LiOr in the vicinity of this line. The trace obtained from the false alarm point association is discarded. If two tracks belong to the same target, two tracks viAnd vjThe track similarity between them can be expressed as
Wherein,is the midpoint of the trace jAnd a straight line LiThe distance of the straight line.
Space-time constraint: since the vehicle cannot appear in two different locations at the same time, it is not in the same frameCo-located moving objects are not necessarily the same object. A vehicle moving on the road cannot necessarily be associated with an object in the current field of view if it goes out of the range of the satellite shots and if the trajectory segment viAnd vjIf there is an overlapping area or one of the associated tracks already exists, the association between the two tracks is S (v)i,vj)＝0。
From the three constraints, the similarity constraint between two trajectory segments can be obtained as follows:
S(vi,vj)＝w1Sm(vi,vj)+w2St(vi,vj) (6)
wherein S ism(. and S)t(. cndot.) represents motion similarity and trajectory similarity, respectively. w is a1And w2Is its weighting factor.
And after the similarity among different tracks is obtained, the candidate target tracks are formed by utilizing the Hungarian algorithm in a correlation mode, and the tracks after the correlation are not straight lines due to the fact that detection points generated by false alarms are ignored, so that straight line fitting is carried out on the generated tracks. And counting the number of points on the straight line according to the best matching degree, namely the matching degree of the inner point rate which is in line with the obtained linear equation in the track and the target track set and the calculated model, and keeping the parameters of the linear equation with the best matching degree. In the invention, when the track length exceeds TlengthAnd the inner point rate is higher than TinnerThen, the moving object on the track is considered as the real object.
And for the candidate moving target, using the motion similarity constraint of hierarchical track association and high-level track association, and selecting tau as 5 frames to reduce the computational complexity. Motion similarity and trajectory similarity weights w1And w20.8 and 0.2, respectively. Fitting a straight line according to the obtained track when the track length T islengthMore than or equal to 10 and fitting the point rate T in the straight lineinnerWhen the distance is higher than 0.6, the detection point on the track is considered as the actual target detection pointAnd outputting the detected weak motion target point in the image.
Therefore, weak and small moving objects in the satellite images are obtained, the moving objects which are filtered based on the space-time information false alarm are used as real moving objects, and the target results are output in the images.
Claims (1)
1. A satellite image ground weak and small moving target detection method based on target saliency is characterized by comprising the following steps:
step one, calculating a saliency image for a current input image; recording a certain pixel value in the image as IkWhere k represents the kth pixel in the image, pixel k; let W be a sliding window centered at pixel k; | l | · | | denotes an inter-pixel gray scale distance matrix, doRepresents the distance of pixel k to pixel o; the local saliency of image pixel k is represented as
Wherein f (-) is with respect to doSo that the significance value is smaller as the distance between pixels increases; when the gray difference between a certain pixel and the adjacent pixel is larger, the probability that the certain pixel is a moving target is higher; obtaining the local saliency of each pixel k of the input gray-scale image through an expression (1), and finally forming a saliency image;
secondly, a background image of the input image is obtained by utilizing self-adaptive Gaussian mixture background modeling on the saliency image, and background subtraction is carried out on the background image and the saliency image to obtain a candidate moving target set;
filtering false alarms of the candidate moving target set by adopting space-time information, and inhibiting the false alarms by utilizing target continuity information; generating associations of trajectory segments hierarchically based on the motion, trajectory and spatiotemporal information; dividing the track association into a primary track association and a high-level track association;
primary track association: assuming that m 'moving objects exist in the t-th frame, and n' moving objects exist in the t + 1-th frame; the Euclidean distance between the moving target in the current frame and each moving target in the next frame is used as an input matrix, and the Hungarian algorithm is used for calculating the moving target association information of the t frame and the t +1 frame; if m '< n', namely the number of the motion targets in the t +1 th frame is more, taking the motion target which is not related as a new motion target; if m 'is more than or equal to n', namely the number of moving targets in the t frame is more, keeping the moving targets for two frames, and then judging whether the moving targets are false alarms or not according to subsequent association;
high-level track association: extracting detection points from each frame by using a Hungarian algorithm to form a short track set; ensuring the accuracy of the long track by utilizing motion similarity constraint, track similarity constraint and space-time constraint;
motion similarity constraint: linear function ofA function representing the distance of movement of the object,representing a trajectory viCoordinates of a position corresponding to the mth frame, Δ t represents a time interval, and v represents a velocity constant in the time interval; order toRepresenting a trajectory viInAndthe time interval in between; temporally adjacent tracks viAnd a track vjBackward and forward prediction respectively, to reduce the computational complexity, only tau frame is calculated, and then the track viAnd vjThe motion similarities between are:
wherein dis1And dis2Respectively as follows:
and (3) constraint of track similarity: a vehicle moving on a road, the track of which is a straight line in a short time domain; therefore, a straight line is fitted according to the short track by using the RANSAC algorithm; trajectory viShould be distributed over the trajectory viEstimated straight line LiOn or near this line; if a certain track is obtained from the false alarm point, discarding the track; two tracks viAnd vjThe track similarity between them is expressed as
Wherein,is a trajectory viMidpointAnd a straight line LjThe linear distance of (d);
space-time constraint: since the vehicle cannot appear at two different locations at the same time, moving objects at different locations in the same frame are necessarily not the same object; a vehicle moving on a road must not be associated with an object in the current field of view if it goes out of range of the satellite photograph, and the trajectory viAnd vjIf there is an overlapping area or one of the associated tracks already exists, the association between the two tracks is S (v)i,vj)＝0；
According to the motion similarity constraint, the track similarity constraint and the space-time constraint, the similarity constraint between the two tracks is obtained as follows:
S(vi,vj)＝w1Smotion(vi,vj)+w2Strack(vi,vj) (6)
wherein S ismotion(. and S)track() represents motion similarity and trajectory similarity, respectively; w is a1And w2Is a weight factor;
after the similarities among different tracks are obtained, the candidate target tracks are formed by utilizing the Hungary algorithm in a correlation mode, the tracks after correlation are not straight lines due to the fact that detection points generated by false alarms are ignored, and therefore straight line fitting is conducted on the generated tracks; according to the inner point rate of the linear equation in the track and the matching degree of the candidate target track and the straight line obtained by fitting, the linear equation with the best matching degree is reservedThe parameters of (1); when the track length exceeds TlengthAnd the inner point rate is higher than TinnerWhen the target is in the track, the moving target on the track is considered to be a real moving target;
therefore, weak and small moving objects in the satellite images are obtained, the moving objects which are filtered based on the space-time information false alarm are used as real moving objects, and moving object results are output in the images.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610091741.4A CN105654516B (en) | 2016-02-18 | 2016-02-18 | Satellite image based on target conspicuousness is to ground weak moving target detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610091741.4A CN105654516B (en) | 2016-02-18 | 2016-02-18 | Satellite image based on target conspicuousness is to ground weak moving target detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105654516A CN105654516A (en) | 2016-06-08 |
| CN105654516B true CN105654516B (en) | 2019-03-26 |
Family
ID=56488552
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610091741.4A Active CN105654516B (en) | 2016-02-18 | 2016-02-18 | Satellite image based on target conspicuousness is to ground weak moving target detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105654516B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10002109B2 (en) * | 2016-06-24 | 2018-06-19 | Freeport-Mcmoran Inc. | Systems and methods of correlating satellite position data with terrestrial features |
| CN107895384A (en) * | 2017-12-01 | 2018-04-10 | 中国科学院长春光学精密机械与物理研究所 | Target extraction method and device |
| CN109712169A (en) * | 2018-11-15 | 2019-05-03 | 上海卫星工程研究所 | Moving-target motion profile prediction technique and method for autonomous tracking based on EO-1 hyperion |
| CN110057449B (en) * | 2019-04-29 | 2021-06-18 | 上海卫星工程研究所 | Air target height and speed calculation method based on field-of-view spectroscopic hyperspectral detection |
| CN110084231A (en) * | 2019-04-29 | 2019-08-02 | 北京富吉瑞光电科技有限公司 | A kind of method and apparatus of infrared panoramic system multi-target detection |
| CN110502968B (en) * | 2019-07-01 | 2022-03-25 | 西安理工大学 | Method for detecting infrared small and weak moving target based on track point space-time consistency |
| CN111862006A (en) * | 2020-06-30 | 2020-10-30 | 北京北方智图信息技术有限公司 | Detection method and device for small aircraft |
| CN112818905B (en) * | 2020-12-10 | 2022-08-26 | 北京航空航天大学 | Finite pixel vehicle target detection method based on attention and spatio-temporal information |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101344967A (en) * | 2008-09-02 | 2009-01-14 | 西北工业大学 | Detection method for small mobile objective in astronomical image |
| CN101520892A (en) * | 2009-03-17 | 2009-09-02 | 西北工业大学 | Detection method of small objects in visible light image |
| CN101546429A (en) * | 2009-05-05 | 2009-09-30 | 西北工业大学 | Detection method of small target of feature space |
| CN101567087A (en) * | 2009-05-25 | 2009-10-28 | 北京航空航天大学 | Method for detecting and tracking small and weak target of infrared sequence image under complex sky background |
| CN102081800A (en) * | 2011-01-06 | 2011-06-01 | 西北工业大学 | Method for detecting spatial weak moving target |
| CN102096829A (en) * | 2011-01-06 | 2011-06-15 | 西北工业大学 | Iterative optimization distance categorization-based space weak and small target detection method |
| CN102103748A (en) * | 2010-12-14 | 2011-06-22 | 西北工业大学 | Method for detecting and tracking infrared small target in complex background |
| CN104143197A (en) * | 2014-07-31 | 2014-11-12 | 中国科学院自动化研究所 | Detection method for moving vehicles in aerial photography scene |
| CN104200492A (en) * | 2014-08-25 | 2014-12-10 | 西北工业大学 | Automatic detecting and tracking method for aerial video target based on trajectory constraint |
| CN105261043A (en) * | 2015-11-20 | 2016-01-20 | 天津大学 | Video motion object detection method on the basis of significance detection |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201001338A (en) * | 2008-06-16 | 2010-01-01 | Huper Lab Co Ltd | Method of detecting moving objects |
-
2016
- 2016-02-18 CN CN201610091741.4A patent/CN105654516B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101344967A (en) * | 2008-09-02 | 2009-01-14 | 西北工业大学 | Detection method for small mobile objective in astronomical image |
| CN101520892A (en) * | 2009-03-17 | 2009-09-02 | 西北工业大学 | Detection method of small objects in visible light image |
| CN101546429A (en) * | 2009-05-05 | 2009-09-30 | 西北工业大学 | Detection method of small target of feature space |
| CN101567087A (en) * | 2009-05-25 | 2009-10-28 | 北京航空航天大学 | Method for detecting and tracking small and weak target of infrared sequence image under complex sky background |
| CN102103748A (en) * | 2010-12-14 | 2011-06-22 | 西北工业大学 | Method for detecting and tracking infrared small target in complex background |
| CN102081800A (en) * | 2011-01-06 | 2011-06-01 | 西北工业大学 | Method for detecting spatial weak moving target |
| CN102096829A (en) * | 2011-01-06 | 2011-06-15 | 西北工业大学 | Iterative optimization distance categorization-based space weak and small target detection method |
| CN104143197A (en) * | 2014-07-31 | 2014-11-12 | 中国科学院自动化研究所 | Detection method for moving vehicles in aerial photography scene |
| CN104200492A (en) * | 2014-08-25 | 2014-12-10 | 西北工业大学 | Automatic detecting and tracking method for aerial video target based on trajectory constraint |
| CN105261043A (en) * | 2015-11-20 | 2016-01-20 | 天津大学 | Video motion object detection method on the basis of significance detection |
Non-Patent Citations (5)
| Title |
|---|
| AN EFFECTIVE DETECTION ALGORITHM FOR MOVING OBJECT WITH COMPLEX BACKGROUND;FU-YUAN HU et al.;《Proceedings of the Fourth International Conference on Machine Learning and Cybernetics》;20050831;5011-5015 |
| Small and Dim Moving Target Detection in Deep Space Background;Jinqiu Sun et al.;《MMSP 2008》;20081231;902-905 |
| Vehicle Detection in Very High Resolution Satellite Images of City Areas;Jens Leitloff et al.;《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》;20100731;第48卷(第7期);2795-2806 |
| Visual Attention Detection in Video Sequences Using Spatiotemporal Cues;Yun Zhai et al.;《Proceedings of the 14th ACM international conference on Multimedia》;20061231;815-824 |
| 基于迭代距离分类与轨迹关联检测空间弱小目标;姚睿 等;《光学精密工程》;20120131;第20卷(第1期);179-189 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105654516A (en) | 2016-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105654516B (en) | Satellite image based on target conspicuousness is to ground weak moving target detection method | |
| CN112669349B (en) | Passenger flow statistics method, electronic equipment and storage medium | |
| CN109242884B (en) | Remote sensing video target tracking method based on JCFNet network | |
| Benfold et al. | Guiding visual surveillance by tracking human attention. | |
| US9213901B2 (en) | Robust and computationally efficient video-based object tracking in regularized motion environments | |
| CN106778712B (en) | Multi-target detection and tracking method | |
| Cheng et al. | Intelligent highway traffic surveillance with self-diagnosis abilities | |
| US10896495B2 (en) | Method for detecting and tracking target object, target object tracking apparatus, and computer-program product | |
| US10685212B2 (en) | Three dimensional (3d) object detection | |
| CN106127807A (en) | A kind of real-time video multiclass multi-object tracking method | |
| CN110555868A (en) | method for detecting small moving target under complex ground background | |
| CN112862845A (en) | Lane line reconstruction method and device based on confidence evaluation | |
| Liu et al. | Multi-type road marking recognition using adaboost detection and extreme learning machine classification | |
| CN113129336A (en) | End-to-end multi-vehicle tracking method, system and computer readable medium | |
| Dhulavvagol et al. | Vehical tracking and speed estimation of moving vehicles for traffic surveillance applications | |
| Zhao et al. | APPOS: An adaptive partial occlusion segmentation method for multiple vehicles tracking | |
| Chen et al. | A novel AMS-DAT algorithm for moving vehicle detection in a satellite video | |
| CN111275733A (en) | Method for realizing rapid tracking processing of multiple ships based on deep learning target detection technology | |
| CN111681262A (en) | Method for detecting infrared dim target under complex background based on neighborhood gradient | |
| Kanhere et al. | Real-time detection and tracking of vehicle base fronts for measuring traffic counts and speeds on highways | |
| Balisavira et al. | Real-time object detection by road plane segmentation technique for ADAS | |
| Muril et al. | A review on deep learning and nondeep learning approach for lane detection system | |
| CN107256382A (en) | Virtual bumper control method and system based on image recognition | |
| Kodeeswari et al. | Lane line detection in real time based on morphological operations for driver assistance system | |
| Płaczek | A real time vehicle detection algorithm for vision-based sensors |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |