JP2015169963A - Object detection system and object detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an object detection system capable of increasing detection accuracy of an object while operating the system, and an object detection method.SOLUTION: An acquisition unit acquires an image obtained by photographing an object body. A parameter calculation unit detects a component configuring an object included in the image acquired by the acquisition unit and calculates a geometric parameter of the component. An object detection unit detects a type of the object on the basis of the comparison between the geometric parameter calculated by the parameter calculation unit and a reference parameter set for each type of the object. An information management unit accumulates at least the geometric parameter calculated by the parameter calculation unit in a storage unit for each type of the object detected by the object detection unit. A display control unit causes a display unit to display an image based on a distribution of the geometric parameters accumulated in the storage unit.

Description

  Embodiments described herein relate generally to an object detection system and an object detection method.

  By analyzing an image of an object, various objects described in the object (marks, identification shapes, character strings, patterns and other visible geometric features), or objects formed by the structure of the object itself Techniques for detecting are known. When an object is detected, a process of comparing the captured image itself or the feature amount of the captured image with a reference image or data is performed. The reference image or data may be corrected by feeding back the object detection result. However, in the conventional technique, the object detection accuracy may not be improved.

Japanese Patent No. 3447280

  The problem to be solved by the present invention is to provide an object detection system and an object detection method capable of improving object detection accuracy while operating the system.

  The object detection system according to the embodiment includes an acquisition unit, a parameter calculation unit, an object detection unit, an information management unit, and a display control unit. The acquisition unit acquires an image obtained by imaging an object. The parameter calculation unit detects a component constituting an object included in the image acquired by the acquisition unit, and calculates a geometric parameter of the component. The object detection unit detects the type of the object based on a comparison between the geometric parameter calculated by the parameter calculation unit and a reference parameter set for each type of object. The information management unit accumulates at least the geometric parameters calculated by the parameter calculation unit in the storage unit for each type of object detected by the object detection unit. The display control unit causes the display unit to display an image based on the distribution of the geometric parameters accumulated in the storage unit.

It is a figure showing an example of composition of object detection system 1 of an embodiment. It is a figure which shows the other example of a structure of the object detection system 1 of embodiment. It is the figure which illustrated object OB which is a fee amount stamp stamped on delivery thing D. It is a figure which shows an example of the component which comprises the object OB. It is the figure which illustrated the geometric parameter of each component which the component detection / parameter calculation part 50 calculates. It is the figure which illustrated the geometric parameter of each component which the component detection / parameter calculation part 50 calculates. It is the figure which illustrated the geometric parameter of each component which the component detection / parameter calculation part 50 calculates. 6 is a diagram illustrating an example of reference parameters stored in a reference parameter 34. FIG. It is a flowchart which shows an example of the flow of a process in an object detection stage. It is the figure which illustrated display image IM1 containing the histogram based on the appearance frequency of a geometric parameter. It is the figure which illustrated display image IM2 on which the histogram based on the appearance frequency of a plurality of geometric parameters was superimposed. It is the figure which illustrated image IM3 expressing the appearance frequency of the geometric parameter. It is the figure which illustrated display image IM4 which superimposed the histogram based on the appearance frequency of a geometric parameter, and the upper limit UL and the lower limit LL of the reference parameter corresponding to the geometric parameter. A display image IM5 in which an image representing the appearance frequency of a geometric parameter and a composite image corresponding to an upper limit value UL and a lower limit value LL of the reference parameter corresponding to the geometric parameter are superimposed by shading or color change. FIG. It is the figure which illustrated display image IM6 containing the histogram based on the appearance frequency of the detection score for every some component of object OB. It is the figure which illustrated object OB containing a plurality of components. It is a flowchart which shows an example of the flow of a process in a detection result analysis stage.

  Hereinafter, an object detection system and an object detection method of an embodiment will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating a configuration example of an object detection system 1 according to the embodiment. For example, the object detection system 1 acquires an image captured by the camera CA of a delivery letter D such as a postal letter or a parcel conveyed by the conveyance mechanism CV, and displays a fee stamp (indicating payment of the fee) included in the image. This is a system for automatically detecting an object OB such as a mark). Note that the camera CA may be a camera that captures a certain plane area, or may be a line sensor.

  The object detection system 1 includes, for example, an input unit 10, a display unit 20, a storage unit 30, an image input unit 40, a component element detection / parameter calculation unit 50, an object detection unit 52, and a result output unit 54. The information management unit 56 and the display control / analysis unit 58 are provided.

  The input unit 10 is an input device such as a keyboard, a mouse, or a touch panel, for example. The display unit 20 is, for example, an LCD (Liquid Crystal Display) or an organic EL (Electroluminescence) display device. The storage unit 30 is a storage device such as an HDD (Hard Disk Drive), a flash memory, or a RAM (Random Access Memory). The storage unit 30 includes a parameter storage unit 32 and a reference parameter storage unit 34. The image input unit 40 includes a network card for connecting to a network connecting the camera CA and the object detection system 1. Note that the image input unit 40 may be connected to an electronic medium in which image data is stored and may input a detection target image that has been captured in advance.

  For example, the component detection / parameter calculation unit 50, the object detection unit 52, the result output unit 54, the information management unit 56, and the display control / analysis unit 58 are executed by a processor (not shown) of the object detection system 1 executing a program. It is a functioning software function unit. Some or all of these functional units may be hardware functional units such as LSI (Large Scale Integration) and ASIC (Application Specific Integrated Circuit).

  The object detection system 1 may include two or more devices. FIG. 2 is a diagram illustrating another example of the configuration of the object detection system 1 according to the embodiment. As shown in the figure, the object detection system 1 includes an object detection device 2 having an image input unit (acquisition unit) 40, a component element detection / parameter calculation unit 50, an object detection unit 52, and a result output unit 54, and an input unit 10. The detection result analysis apparatus 3 having the display control / analysis unit 48 may be included. The object detection device 2 and the detection result analysis device 3 are connected by a LAN (Local Area Network) or the like. In the configuration of FIG. 2, the storage unit 30 and the information management unit 56 may be provided on either side of the object detection device 2 and the detection result analysis device 3. What are the object detection device 2 and the detection result analysis device 3? It may be a separate one.

[Object detection processing]
Hereinafter, the contents of the object detection process executed by the object detection system 1 will be described. The component detection / parameter calculation unit 50 calculates the geometric parameters of the object OB included in the image acquired by the image input unit 40. FIG. 3 is a diagram illustrating an object OB (objects OB1, OB2) that is a fee stamp stamped on the delivery item D. As shown in FIG. As shown in the figure, the object OB is a combination of a specific figure (circle, ellipse, rectangle, etc.), a character string, and the like. Further, the object OB may include a wavy line pattern simulating the edge of a stamp, or a rectangle or pattern surrounded by the wavy line.
The component detection / parameter calculation unit 50 detects the object OB as a combination of a plurality of components and calculates a geometric parameter for each component. The geometric parameter is information represented by a numerical value. FIG. 4 is a diagram illustrating an example of components that constitute the object OB. As illustrated, the component detection / parameter calculation unit 50 detects a circle OBC and a character string OBL as components for an object OB1 including a character string in a circle. In addition, the component detection / parameter calculation unit 50 detects the rectangle OBS and the character string OBL as the component for the object OB2 including the character string in the rectangle.

  5 to 7 are diagrams illustrating examples of geometric parameters of each component calculated by the component detection / parameter calculation unit 50. As shown in FIG. 5, the component detection / parameter calculation unit 50 calculates the radius OBC1 for the circle OBC and the lengths OBS1 and OBS2 of each side for the rectangular OBS as geometric parameters. Further, as shown in FIG. 6, the component element detection / parameter calculation unit 50, for the character string OBL, the length OBL1 of the character string, the height (width in the case of vertical writing) OBL2, the width of one character (Height in case of vertical writing) OBL3 and character spacing OBL4 are calculated as geometric parameters. As shown in FIG. 6, the component element detection / parameter calculation unit 50 determines the arrangement of the character string OBL in the rectangular OBS including the character string OBL (distances OBL5 and OBL6 from the left end and the upper end) as geometric parameters. May be calculated as In addition, the component detection / parameter calculation unit 50 calculates distances OBD1 and OBD2 from the left end and the upper end of the delivery object D for the object OB as geometric parameters. In addition to these, when the component is a wavy line, the component detection / parameter calculation unit 50 may calculate the pitch, phase, angle, distance from a certain point, and the like as geometric parameters, A value indicating a relative relationship between geometric parameters such as a length ratio between one side and another side may be calculated as a geometric parameter.

  Here, a method of detecting the constituent elements of the object OB by the constituent element detection / parameter calculation unit 50 will be described. When the component is a circle OBC, the component element detection / parameter calculation unit 50 obtains the center coordinates and the radius of the circle by voting processing, for example, by Hough transform. Further, when the component is a rectangular OBS, the component detection / parameter calculation unit 50 detects the angle θ of the straight line and the distance ρ from the origin by the Hough transform, and then calculates the angle, the end point position, the distance between the straight lines, and the like. Based on the four sides of the rectangle, the rectangle detection result is selected. As a method of selecting four sides, for example, two parallel straight lines and another two parallel straight lines perpendicular to the straight line are selected, and the end points of the straight line are on or between the two parallel straight lines perpendicular to the straight line. If this condition is observed for all four straight lines, a method of selecting the straight lines constituting the rectangle is adopted. Further, when the constituent element is a character string OBL, the constituent element detection / parameter calculation unit 50 extracts, for example, a connected region obtained by the labeling process of the binary image as a character candidate, and the character candidates are arranged in the horizontal direction. The part is detected as a character string. In addition, when the constituent element is a wavy line, the constituent element detection / parameter calculation unit 50 approximates the wavy line by a periodic function, and sets the periodic function to two parameters (angle θ, distance ρ from the origin) as a straight line of the axis. The three parameters of the period ω, the amplitude A, and the phase α are added, and the values are determined by voting as in the Hough transform. In this case, the period and amplitude of the wavy line to be detected may be fixed, and voting may be limited to three parameters: angle θ, distance ρ from the origin, and phase α. Note that the component element detection / parameter calculation unit 50 performs processing with a detection target limited in advance based on a reference parameter stored in a reference parameter storage unit 34 (to be described later) in order to increase the processing speed and efficiency. Also good.

  In addition, the component element detection / parameter calculation unit 50 may calculate a detection score indicating the reliability of detection of the component element of the object OB. For example, when the constituent element is a circle OBC or a rectangular OBS, the constituent element detection / parameter calculation unit 50 calculates a ratio of black pixels (pixels whose luminance is less than the threshold) in pixels on an arc or a rectangular straight line as a detection score. Calculate as In addition, the component element detection / parameter calculation unit 50 may also use the number and ratio of white pixels (pixels whose luminance is equal to or greater than a threshold) in pixels in the vicinity of an arc or a rectangular straight line. When the constituent element is the character string OBL, the constituent element detection / parameter calculation unit 50 calculates the average of the character recognition scores of the characters as the detection score. Here, in the case of a specific character string such as “postpaid postage” with a fee stamp, a character recognition score when recognized as a corresponding word may be used. In addition, when it is determined to be a number like a fee, a character recognition score when it is recognized as any number may be used.

  The object detection unit 52 detects the object OB by comparing the geometric parameter calculated by the component element detection / parameter calculation unit 50 with the reference parameter stored in the reference parameter storage unit 34. Here, the reference parameter is a value set for each object OB and further for each component constituting the object OB, and includes, for example, an upper limit value, a lower limit value, and a standard value. FIG. 8 is a diagram illustrating an example of the reference parameter stored in the reference parameter 34.

  For example, when all of the geometric parameters calculated by the component element detection / parameter calculation unit 50 are within the upper limit value and the lower limit value of the reference parameter, the object detection unit 52 displays the corresponding object OB in the image. It is determined that it is included. Further, even if some geometric parameters do not fit between the upper limit value and the lower limit value of the reference parameter, the object detection unit 52 is applicable when other geometric parameters are close to the standard value. It may be determined that the object OB is included in the image. As described above, the object detection unit 52 comprehensively determines whether or not the object OB is included in the image based on the result of comparison with the reference parameter for each of the plurality of components.

  Further, the object detection unit 52 may determine whether or not the corresponding object OB is included in the image using the detection score calculated by the component element detection / parameter calculation unit 50. In this case, the object detection unit 52 weights the detection score according to the nature of the detection score (for example, whether the maximum value is 100 or 200), and sums the weighted sums as the reference value. It may be determined whether or not the corresponding object OB is included in the image. In addition, when the frequency of simultaneous appearance of two or more specific figures (components) having specific parameter values included in the object OB to be detected is low, the object detection unit 52 simultaneously displays the two figures. If detected, the detection score may be weighted higher.

  The result output unit 54 outputs the detection result obtained by the object detection unit 52. The output destination of the detection result is, for example, a delivery device D for sorting deliveries D (not shown). The sorting apparatus refers to the detection result by the object detection unit 52 and performs various processes.

  The information management unit 56 accumulates at least the geometric parameters calculated by the component element detection / parameter calculation unit 50 in the parameter accumulation unit 32 of the storage unit 30 for each type of object OB detected by the object detection unit 52. . Further, the information management unit 56 may store the detection score calculated by the component element detection / parameter calculation unit 50 in the parameter storage unit 32 in addition to the geometric parameters. Further, even when the type of the object OB is not specified by the object detection unit 52, the information management unit 56 associates the most probable object OB with the geometric parameter calculated by the component element detection / parameter calculation unit 50. Or the like may be stored in the parameter storage unit 32. The function of the information management unit 56 may be embedded in the component element detection / parameter calculation unit 50 or the object detection unit 52.

  Here, in principle, the information management unit 56 causes the storage unit 30 to store only digitized information such as geometric parameters and detection scores. Accordingly, the storage capacity of the storage unit 30 and the transfer time required for storing the image can be saved as compared with the case where the image input from the camera CA is stored in the storage unit 30.

  FIG. 9 is a flowchart showing an example of the processing flow in the object detection stage described above. First, the processor of the object detection system 1 reads a reference parameter from the reference parameter storage unit 34 (step S100). Next, the image input unit 40 inputs an image from the camera CA (step S102), the component detection / parameter calculation unit 50 detects the component (step S104), and calculates a parameter for the detected component. (Step S106). Then, the object detection unit 52 determines whether or not the object OB is included in the image (object determination; step S108), and the result output unit 54 outputs the detection result by the object detection unit 52 (step S110). . The processes in steps S102 to S110 are executed until the input of the image is completed (step S112).

[Detection result analysis processing]
Here, the object OB imaged by the camera CA may have various types, and the size and shape may vary. For this reason, if the reference parameter given as the initial value is continuously used as it is, the possibility of erroneous detection or detection omission cannot be excluded. Therefore, in the object detection system 1 of the embodiment, an image based on the distribution of the geometric parameters stored in the parameter storage unit 32 of the storage unit 30 is displayed on the display unit 20 to show to the adjuster, or the parameter storage unit 32 Based on the distribution of the geometric parameters accumulated in, the adjustment parameter is urged to be adjusted by the adjuster, or the reference parameter is automatically corrected. The correction of the reference parameter can be performed by the adjuster performing an input operation on the input unit 10. Hereinafter, such detection result analysis processing will be described.

  The display control / analysis unit 58 causes the display unit 20 to display an image based on the geometric parameter distribution stored in the parameter storage unit 32 of the storage unit 30. For example, the display control / analysis unit 58 causes the display unit 20 to display an image including a histogram based on the appearance frequency of the geometric parameter for one object OB. FIG. 10 is a diagram illustrating a display image IM1 including a histogram based on the appearance frequency of geometric parameters. Here, the appearance frequency of the geometric parameter may be limited to a component having a detection score of a certain value or more, or may be weighted by a detection score. In this way, the adjuster who has viewed the display image IM1 can grasp the tendency of the distribution of the geometric parameter detected.

  Further, the display control / analysis unit 58 may superimpose and display an image in which histograms of a plurality of geometric parameters are superimposed on the display unit 20. FIG. 11 is a diagram illustrating a display image IM2 on which a histogram based on the appearance frequency of a plurality of geometric parameters is superimposed. In this case, the display control / analysis unit 58 may display a plurality of geometric parameters on a graph after normalizing them so as to facilitate comparison. In this way, the adjuster who has viewed the display image IM2 can easily grasp the difference in the distribution tendency between the geometric parameters.

  In addition, the display control / analysis unit 58 may cause the display unit 20 to display an image expressing the appearance frequency of the geometric parameter by the change in shade or color. FIG. 12 is a diagram illustrating an image IM3 representing the appearance frequency of the geometric parameter. In the figure, OB * is a composite image expressing standard constituent elements, and G is a portion expressing the appearance frequency of geometric parameters by shading or color change. In this way, the adjuster who has viewed the display image IM3 can intuitively grasp the distribution of the geometric parameters.

  The display control / analysis unit 58 superimposes the histogram based on the appearance frequency of the geometric parameter on the display unit 20 by superimposing the upper limit value UL and the lower limit value LL of the reference parameter corresponding to the geometric parameter. You may let them. FIG. 13 is a diagram illustrating a display image IM4 in which a histogram based on the appearance frequency of a geometric parameter and an upper limit value UL and a lower limit value LL of a reference parameter corresponding to the geometric parameter are superimposed. In this way, the adjuster who viewed the display image IM4 can grasp whether or not the reference parameter and the distribution of the geometric parameter are properly matched.

  In addition, the display control / analysis unit 58 combines an image representing the appearance frequency of a geometric parameter based on shading or color change and a reference parameter upper limit value UL and lower limit value LL corresponding to the geometric parameter. An image may be superimposed and displayed on the display unit 20 in a superimposed manner. In FIG. 14, an image representing the appearance frequency of a geometric parameter is superimposed with a composite image corresponding to the upper limit value UL and the lower limit value LL of the reference parameter corresponding to the geometric parameter by shading or color change. It is the figure which illustrated display image IM5. In this way, the adjuster who has viewed the display image IM5 can intuitively understand whether or not the reference parameter and the distribution of the geometric parameter are properly matched.

  Further, the display control / analysis unit 58 may cause the display unit 20 to display an image based on the detection score for each component of the object OB instead of the geometric parameter itself. FIG. 15 is a diagram illustrating a display image IM6 including a histogram based on the appearance frequency of the detection score for each of the plurality of components of the object OB. In this case, the display control / analysis unit 58 may display a plurality of detection scores on a graph after normalizing them so as to facilitate comparison. In this way, the adjuster who sees the display image IM6 can grasp what detection score each component detects, and can adjust the detection logic as necessary.

  The display control / analysis unit 58 may control the display unit 20 so as to provide a user interface for switching the various images exemplified above. The user interface, for example, superimposes on the image of the object OB to be detected, shows the geometric parameters related thereto, and when the user clicks a symbol or area indicating the geometric parameters, the appearance frequency histogram of the geometric parameters Is displayed.

  Hereinafter, analysis processing by the display control / analysis unit 58 will be described. The display control / analysis unit 58 causes the display unit 20 to display a warning display when the geometric parameter distribution accumulated in the parameter accumulation unit 32 and the reference parameter are different. “The geometric parameter distribution is different from the reference parameter” means, for example, that a predetermined ratio (for example, 95%) of the distribution does not fall between the upper limit value and the lower limit value of the reference parameter. Say. In addition, when the geometric parameter average, mode, and median are more than a certain percentage away from the standard value of the standard parameter, the geometric parameter distribution and standard parameter are different. It can be true. As described above, the object detection unit 52 comprehensively determines whether or not the object OB is included in the image based on the result of comparison with the reference parameter for each of the plurality of components. Therefore, in general, even when it is determined that the object OB, it may include a component that deviates from the reference parameter. The display control / analysis unit 58 performs analysis processing as described above to find that some of the components deviate from the reference parameter and notify the adjuster or automatically correct the reference parameter. To do. For example, the object shown in FIG. 16 includes two circles OBC1, OBC2, a rectangle OBS1, and three character strings OBL1, OBL2, OBL3. Of these, the detection frequency of the outer circle OBC1 is the inner When the detection frequency of the circle OBC2 and the rectangular OBS1 is extremely low, the outer circle OBC1 may be inappropriate as a figure used for detecting the object OB, and a warning is displayed to the adjuster. .

  In addition, the display control / analysis unit 58 determines whether or not the weight of the detection score used when the object detection unit 52 comprehensively determines is appropriate, notifies the adjuster of the determination result, or sets a weighting coefficient. It may be corrected automatically. For example, the display control / analysis unit 58 may detect that the detection score weighted by the weighting coefficient of the constituent elements constituting the object OB is relatively high and is detected alone, that is, configures the object OB. When the weighted detection score of a component having a small number takes a relatively low value, it is determined that the appearance frequency of the detection score matches the weighting coefficient of the detection score.

  The display control / analysis unit 58 controls the display unit 20 to provide a user interface for adjusting the reference parameter to the adjuster. As an input method, there are a method of inputting a numerical value, an input using a GUI screen such as a slider or a button, and a method of directly dragging a symbol indicating a parameter value setting range on a displayed screen with a mouse or the like.

  In the correction process of the reference parameter, the display control / analysis unit 58 first performs an object OB having a specific geometric parameter as well as the geometric parameter and the detection score for a certain period of time by an operation of the adjuster or automatically. The information management unit 56 is instructed to shift to a mode for saving the original image (image saving mode). The coordinator checks the stored original image and determines whether the object OB having a geometric parameter with a high appearance frequency is the same as or different from the object OB to be detected. If they are the same, the adjuster corrects the reference parameter in accordance with the appearance frequency of the detected geometric parameter. If they are different, the coordinator newly registers a reference parameter as a component of a different object OB.

  On the other hand, if an object OB that should not be detected is captured in the original image, for example, a graphic such as a company logo, it is placed in the reject list in order to avoid subsequent erroneous detection. The reject list is a list of objects OB that are not detected, and is data in which a combination of components and a range of values of geometric parameters are described in the same format as the reference parameter. According to this list, for example, if the fee stamp to be detected is composed of two circles and one rectangle, and the company logo not to be detected is composed of two circles and two rectangles of the same size as the fee stamp, By registering a combination of two circles and two rectangles in the reject list, if there is only one rectangle, it will be detected as a fee stamp, and if there are two rectangles, it will not be detected, thus preventing false detection. it can.

  In addition, after the display control / analysis unit 58 shifts to the image storage mode and stores the original image in the same manner as described above, the object OB having the geometric parameter value having a high appearance frequency is the same as the object OB to be detected. Or different ones may be automatically determined by image processing such as template matching. When it is determined that the objects are the same object OB, the reference parameter is corrected in accordance with the appearance frequency of the detected geometric parameter of the constituent element. If it is not a different object OB or an object OB to be detected, a warning is displayed on the display unit 20.

FIG. 17 is a flowchart showing an example of the processing flow in the detection result analysis stage described above. First, the processor of the object detection system 1 reads a reference parameter from the reference parameter storage unit 34 (step S200). Next, the display control / analysis unit 58 reads the information stored in the parameter storage unit 32 (step S202), and analyzes the geometric parameter and the detection score (step S204). And a display image is displayed on the display part 20 in the various aspect demonstrated above (step S206).
Next, the display control / analysis unit 58 receives an input from the adjuster to the input unit 10 (step S208), and updates the reference parameter based on the input content (step S210).

  According to at least one embodiment described above, at least the geometric parameters calculated by the component detection / parameter calculation unit 50 are accumulated in the storage unit 30 for each type of object OB detected by the object detection unit 52. Since the image based on the accumulated geometric parameter distribution is displayed on the display unit 20, the object detection accuracy can be improved while operating the system.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 1 ... Object detection system, 10 ... Input part, 20 ... Display part, 30 ... Memory | storage part, 32 ... Parameter storage part, 34 ... Reference | standard parameter storage part, 40 ... Image input part, 50 ... Component element detection / parameter calculation part, 52 ... Object detection unit, 54 ... Result output unit, 56 ... Information management unit, 58 ... Display control / analysis unit, CA ... Camera, D ... Delivery item, OB ... Object

Claims (9)

An acquisition unit that acquires an image of an object;
A parameter calculation unit for detecting a component constituting an object included in the image acquired by the acquisition unit and calculating a geometric parameter of the component;
An object detection unit for detecting the type of the object based on a comparison between the geometric parameter calculated by the parameter calculation unit and a reference parameter set for each type of object;
For each type of object detected by the object detection unit, an information management unit that accumulates at least a geometric parameter calculated by the parameter calculation unit in a storage unit;
A display control unit that displays on the display unit an image based on the distribution of the geometric parameters accumulated in the storage unit;
An object detection system comprising: The display control unit causes the display unit to display an image including a histogram indicating the distribution of the geometric parameters.
The object detection system according to claim 1. The display control unit causes the display unit to display an image indicating the distribution of the geometric parameter by shading or color change;
The object detection system according to claim 1 or 2. The display control unit is configured to display an image including a histogram indicating the distribution of the geometric parameter, or an image obtained by superimposing an image indicating the distribution of the geometric parameter with shading or color change and an image based on the reference parameter. , Display on the display unit,
The object detection system according to any one of claims 1 to 3. The parameter calculation unit outputs a detection score when the component is detected,
The display control unit causes the display unit to display an image based on the distribution of the detection scores.
The object detection system according to any one of claims 1 to 4. The geometric parameter includes any of a radius of a circle, a length of one side of a rectangle, a length and width of a character string, a character size, and a character spacing.
The object detection system according to any one of claims 1 to 5. The display control unit causes the display unit to perform a warning display based on a comparison between the geometric parameter distribution stored in the storage unit and the reference parameter.
The object detection system according to any one of claims 1 to 6. In a predetermined case, the display control unit is configured to switch to a mode in which an image acquired by the acquisition unit is stored in the storage unit in addition to the geometric parameter calculated by the parameter calculation unit. Instruct
After displaying the image stored in the storage unit to the adjuster, accepting an input by the adjuster for updating the reference parameter;
The object detection system according to claim 1. Get an image of the object,
Detecting components constituting an object included in the acquired image;
Calculating geometric parameters of the component;
Detecting the type of the object based on a comparison between the calculated geometric parameter and a reference parameter set for each type of object;
For each type of detected object, accumulate at least the calculated geometric parameters in a storage unit;
An image based on the distribution of geometric parameters stored in the storage unit is displayed on the display unit.
Object detection method.

Images