JP2007287093A - Program, method and device for detecting mobile object - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device for detecting a mobile object, which easily and appropriately responds to a situation when an undetected object is generated. <P>SOLUTION: An image obtaining unit 101 stores the obtained image in an image storage unit 102. A first mobile object detector 103 and a second mobile object detector 104 detect a mobile objects image photographed as a shooting image by different detection methods. A detection control unit 107 operates, at first, the first mobile object detector 103 to detect the image of the mobile object. A display control unit 105 makes a display device 120 display the result of detection. The detection control unit 107 makes the first mobile object detector 103, in accordance with a set of settings, detect the image by changing detection conditions for the image for the mobile object, or the second mobile object detector 104 detect the image, when receiving a notice from an input part 106 that a re-detection of the image for the mobile object has been instructed by an operator. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a moving body detection apparatus that detects a moving body photographed as a subject in a captured image.

  A moving body detection device that detects a moving body is used for traffic event analysis or for assisting movement of elderly and disabled persons. As the moving body detection device, as described in Patent Document 1, for example, an apparatus using an image processing technique, that is, an image of a moving body (moving body image) captured as a subject in a captured image is detected. Things are common.

  Usually, a background subtraction method or an interframe subtraction method is used for moving object detection. In the background subtraction method, a background image (model) of a shooting location is prepared in advance, and a moving body image is detected from a difference between the background image (model) and the shot image (difference in pixel value obtained for each pixel). Is. The inter-frame difference method is to detect a moving body image from a difference between captured images (frames).

  FIG. 10 is a diagram for explaining the moving object detection by the background subtraction method. In the figure, the vertical axis indicates the difference value between the background image and the captured image, that is, the difference between the pixel values (hereinafter referred to as “pixel difference value”), and the horizontal axis indicates the X or Y axis in the captured image. The position on the line is shown. Accordingly, FIG. 10 schematically shows a change depending on the position of the difference value obtained in the line. Note that the change of the pixel difference value on the line is not actually linear as shown in FIG.

  When two moving bodies A and B exist on the line, the pixel difference value at the location where the moving bodies A and B are present increases as shown in FIG. Therefore, a moving body image can be detected by appropriately setting a threshold value of a pixel difference value that should be regarded as a moving body. In the example shown in FIG. 10, by setting the detection level 2 among the detection levels 1 and 2 as the threshold value, both the moving bodies A and B can be detected. Hereinafter, the threshold is referred to as “detection level”.

  The pixel difference value can be either positive or negative. When the absolute value is calculated, the detection level to be set can be set to one. Here, for convenience of explanation, it is assumed that the pixel difference value is obtained as an absolute value unless otherwise specified.

  The actual background image changes depending on the weather and time zone. In order to cope with the change, a plurality of background images as models (hereinafter referred to as “background models” to be distinguished from actual background images) are prepared or estimated in the conventional moving body detection device, Some background models are applied as appropriate.

By optimizing the background model to be applied, the moving body image can be detected with higher accuracy. However, the state of the moving body (color, brightness, etc.) varies depending on the moving body, and its width is wide. For this reason, it is very difficult to set a detection level that can reliably detect all moving objects. Even in the same weather and the same time zone, shooting conditions may vary depending on the moving object. For example, in detecting a vehicle as a moving body, a small vehicle may enter the shadow of a large vehicle such as a bus. In such a situation, the brightness differs between large and small vehicles. The greater the difference in brightness, the higher the possibility that a small vehicle cannot be detected. From these facts, it is a fact that not all moving bodies can be reliably detected. Thereby, it is considered important to be able to cope more easily and appropriately when a moving body cannot be detected.
Japanese Patent No. 3461201 JP-A-6-203157 Japanese Patent Laid-Open No. 11-132740

  It is an object of the present invention to provide a moving body detection apparatus that can more easily and appropriately cope with a moving body that cannot be detected.

Both of the moving body detection apparatuses according to the first and second aspects of the present invention are based on the premise that a moving body photographed as a subject in a photographed image is detected, and each includes the following means.
A moving body detection apparatus according to a first aspect includes an image acquisition unit that acquires a captured image, a moving body detection unit that detects a moving body image present in the captured image acquired by the image acquisition unit, and a moving body detection unit. Display control means for displaying the detection result by the display device together with the captured image, instruction means for instructing execution of redetection of the moving body image, and execution of redetection by the instruction means, Detecting means for changing the detection condition of the moving body image, and causing the moving body detection means to execute the detection of the moving body image again.

  The moving body detecting means detects the moving body image from the difference between the captured image and the background image prepared in advance, and the detection control means detects the moving body image by changing the condition of the difference to be regarded as the moving body image. It is desirable to change the conditions. The re-detection instruction by the instruction unit is performed by designating a moving body image displayed by the display control unit and not detected by the moving body detection unit, and the detection control unit includes a position designated by the instruction unit, It is desirable to change the condition of the difference applied to the vicinity thereof.

  The moving body detecting means detects the moving body image from the difference between the photographed image and the background image prepared in advance, and the redetection instruction by the instruction means is displayed by the display control means. The detection control means compares the designated part with the background image when the part is designated by the instruction means, and temporarily compares the background image based on the comparison result. It is desirable to change the detection condition by changing.

  The moving body detecting means includes first and second moving body detecting means that target black and white images and color images as captured images, respectively, and the detection control means performs re-detection by the instruction means. When instructed, it is desirable to change the detection condition by switching the object for detecting the moving body image from the first moving body detecting means to the second moving body detecting means. The second moving body detection means treats a color image expressed by hue information, saturation information, and brightness information as a captured image, thereby moving the moving body from a difference between the color image and a background image prepared in advance. The image detection is performed, and when the execution of re-detection is instructed by the instruction unit, the detection control unit changes the lightness information of one of the color image and the background image to the lightness information of the other, and the second It is desirable to cause the moving body detection means to re-detect the moving body image. Alternatively, the second moving body detecting means assumes a coordinate system in which a color image expressed by three color information of red, green, and blue is handled as a color image, and each of the three color information is a coordinate axis. It is desirable to detect the moving body image based on the position of the coordinate system represented by the three color information of the color image. In that case, it is desirable to specify the direction of the straight line connecting the origin and the position of the coordinate system and the distance of the straight line, and to detect the moving body image from the difference between the specified direction and the distance. Further, the re-detection instruction by the instruction unit can be performed by designating one of the three color information, and the detection control unit operates the color information designated by the instruction unit to operate the second color information. It is desirable to cause the moving body detection means to detect the moving body image.

  The image acquisition means acquires a moving image as a captured image, and the detection control means changes the detection condition when the instruction means instructs execution of re-detection, and when the instruction is executed. It is desirable that the moving body detection means detect the moving body image retroactively from the display position on the moving image displayed by the display control means.

  The moving body detection apparatus according to the second aspect includes image acquisition means for acquiring a color image represented by three color information of red, green, and blue as captured images, and coordinates having the three color information as coordinate axes, respectively. Assuming a system, the position of the coordinate system represented by the three color information of the color image acquired by the image acquisition means is specified, and the direction of a straight line connecting the specified position and the origin of the coordinate system, And a moving body detecting means for further identifying the distance of the straight line and detecting a moving body image from the difference between the identified direction and the distance.

  The moving object detection method of the present invention is for detecting a moving object photographed as a subject in a photographed image, detects a moving object image present in the photographed image, and displays the detection result. Then, the re-detection for the moving body image that is recognized not to be detected by displaying the detection result is performed by automatically changing the detection condition according to the re-detection instruction.

  The program of the present invention is acquired by an image acquisition function for acquiring a captured image and an image acquisition function on the premise that the program is used by a computer used as a mobile object detection device that detects a mobile object captured as a subject in a captured image. A moving body detection function for detecting a moving body image existing in the captured image, a display control function for displaying a detection result by the moving body detection function on the display device together with the captured image, and a redetection of the moving body image. When an instruction function for instructing execution and execution of re-detection are instructed by the instruction function, the detection condition of the moving object image is changed, and the detection of the moving object image by the moving object detection function is executed again. And a detection control function.

  In the present invention, detection of a moving body image present in a captured image is performed, the detection result is displayed, and redetection for a moving body image that is recognized as not detected by the display of the detection result is performed. The detection condition is automatically changed according to the re-detection instruction by the operator.

  By changing the detection condition, it becomes possible to detect a moving body image that could not be detected. Since the detection condition is automatically changed, the operator can easily detect the moving body image that could not be detected. As a result, even if a mobile object that could not be detected is generated, the mobile object detection device can be made to take an appropriate action more easily, and high convenience can be obtained.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a functional configuration diagram of a moving object detection device according to the present embodiment. The moving body detection device acquires a captured image obtained by shooting with a shooting device, for example, a place where a moving body to be detected moves as a shooting place, and is taken as a subject by moving the place. The moving object is detected. As illustrated in FIG. 1, an image acquisition unit 101 that acquires the captured image, an image storage unit 102 that stores the captured image acquired by the acquisition unit 101, and a subject in the captured image stored in the storage unit 102. First and second moving body detection units 103 and 104 that detect moving body images photographed as images, and detection results of the captured images by the first or second moving body detection unit 103 or 104 In addition, a configuration including a display control unit 105 to be displayed on the display device 120, an input unit 106 for an operator (user) to perform various inputs, and a detection control unit 107 for controlling the units 101 to 106 It has become.

  The image acquisition unit 101 receives a captured image from a terminal device connected to the imaging device, for example, via a communication network, or accesses a recording medium such as an optical disk and reads the recorded image recorded on the recording medium. . The input unit 106 is for inputting a command or data instructed by an operator operating a pointing device such as a mouse. The display device 120 on which the display control unit 105 displays the detection result is connected to or mounted on the moving object detection device. The image storage unit 102 is, for example, a hard disk device or a semiconductor memory. The shooting location where the shooting is performed by the shooting device is, for example, a road on which a vehicle passes.

  The moving body detection apparatus having the above-described configuration is realized by causing a personal computer (PC) to execute a program that performs an operation as described later, for example. The program to be executed by the PC, that is, the CPU mounted thereon, may be stored in a mounted or connected hard disk device, but may be recorded in a recording medium such as an optical disk or a magneto-optical disk. Or you may enable it to receive via a communication network.

  The first moving body detection unit 103 handles a monochrome image as a captured image, and the second moving body detection unit 104 handles a color image as the captured image. As a result, when the photographing apparatus captures a color image, that is, when the image acquisition unit 101 acquires a color image as a captured image, the first moving body detection unit 103 targets a monochrome image converted from the color image. The moving body image is detected.

  The detection control unit 107 causes the first or second moving body detection unit 103 or 104 to detect the moving body image, receives the detection result, and outputs the detection result to the display control unit 105. Thereby, the display control unit 105 causes the display device 120 to display the detection result together with the captured image obtained from the image storage unit 102.

  FIG. 6 is a diagram for explaining a method of displaying the detection result of the moving object image. In the figure, reference numeral 601 denotes a photographed subject image, and reference numeral 602 denotes a mouse cursor. In the present embodiment, what can be detected in the subject image 601 present in the captured image is surrounded by a rectangular frame. Thus, the operator can recognize for each subject image 601 whether or not the detection is possible based on the presence or absence of the frame. In order to display such a frame, for example, each movement detection unit 103, 104 notifies the detection control unit 107 of the center position and the width on the XY axis for each detected moving body image 601. ing.

  Actually, it is very difficult to always reliably detect the subject image 601 in the captured image. For this reason, in the present embodiment, the moving object image 601 is re-detected in accordance with an instruction from the operator. The re-detection is performed by automatically changing the detection condition so that the moving body image 601 that could not be detected can be detected. The re-detection by automatically changing the detection condition is performed in accordance with a designation method designated in advance. Hereinafter, the re-detection method will be specifically described with reference to the explanatory diagrams of FIGS.

  In order to reduce the amount of calculation, the detection of the moving object image 601 is first performed by the first moving object detection unit 103. The first moving object detection unit 103 detects the moving object image 601 from the background difference method, that is, the difference between the applied background model and the captured image. The re-detection is performed by the first or second moving body detection unit 103 or 104 according to the designated content (hereinafter referred to as “re-detection setting content” for convenience). First, the redetection performed by the first moving body detection unit 103 will be described in detail. The background model is stored in the image storage unit 102.

  FIG. 7 is a diagram for explaining a first re-detection method performed by the first moving object detection unit 103. Similar to FIG. 10, the vertical axis represents the difference value between the background model and the captured image (pixel difference value, expressed as “level” in the figure), and the horizontal axis represents the line on the line along the X or Y axis in the captured image. Indicates the position. Accordingly, FIG. 7 shows a change depending on the position of the pixel difference value obtained on the line. As in FIG. 10, the change in the pixel difference value on the line is not linear. The same applies to other cases.

  As the detection level that is the threshold value of the pixel difference value that is regarded as constituting the moving object image 601 increases, the width that is regarded as the moving object image 601 tends to become narrower. The moving body image 601 is detected on the condition that the size is within a range corresponding to the assumed moving body. For this reason, in the first redetection method, the detection level is automatically adjusted to redetect a moving body image 601 having an appropriate size (appropriate width on both XY axes). Reference numeral 111 in FIG. 1 denotes an automatic adjustment unit that performs the automatic adjustment.

  The automatic adjustment of the detection level is performed in a direction that makes the width appropriate according to the width of the moving body image 601 that is currently considered. For example, if the current detection level is V and the width of the moving object image 601 at that level V is too small, the level V is then set to the half level V / 2. If the width is too small even at the level V / 2, the level V / 2 is further reduced to half the level V / 4. On the other hand, when the width becomes too large at the level V / 2, the level V / 2 is then set to 1.5 times the level 3V / 4. In this way, if the width is too large, the level is set to the next half of the current level, and if the width is too small, the level is then set to 1.5 times the current level.

  By performing such automatic adjustment (change of detection conditions), if a detection level appropriate for the moving body image 601 that has not been detected exists, it can be found. For this reason, the moving body image 601 that has not been detected can be recognized by re-detection. The reason why the automatic adjustment is performed in the above-described manner is to find an appropriate detection level more quickly. Redetection while automatically adjusting (changing) the detection level may not be able to properly detect the moving object image 601 by adjusting the detection level, and therefore, for example, the width of the moving object image 601 is appropriate. If the absolute value of the difference between the detection level set next and the detection level set in the past is equal to or less than a predetermined value, the process is terminated.

  When instructing redetection by the first redetection method, a moving body image 601 that has not been detected is designated. The designation is performed by a click operation with the mouse cursor 602 moved to the corresponding position. The detection control unit 107 notified from the input unit 106 that such a click operation has been performed notifies the first moving body detection unit 103 of the position where the click operation has been performed, and performs re-detection. Instruct. Accordingly, the first moving body detection unit 103 detects the moving body image 601 while performing automatic adjustment by the automatic adjustment unit 111.

  FIG. 8 is a diagram for explaining a second redetection method performed by the first moving body detection unit 103. In FIG. 8, the vertical axis indicates the pixel value (luminance value) of the captured image, that is, the pixel difference value based on the applied background model, and the horizontal axis indicates the position on the line along the X or Y axis in the captured image. Is shown. Accordingly, FIG. 8 shows a change due to the position of the pixel difference value obtained on the line.

  The moving body image 601 changes depending on the shooting situation. Therefore, a plurality of background models are prepared, or the models are estimated appropriately. Typical causes for adapting the background model include background changes due to weather and time zones, and background object changes such as tree shaking.

  There may be a case where shooting conditions differ between moving bodies. For example, a small vehicle may enter the shadow of a large vehicle such as a bus. In such a situation, large vehicles are bright and small vehicles are dark. As a result, the appropriate background model for large and small vehicles may be different. Therefore, in the second re-detection method, the background model is compared with the background portion of the captured image, and the background model is temporarily changed according to the comparison result. The reason for changing the background model is that the amount of calculation can be further suppressed when the background model is changed. By making the background model more appropriate by the change, the moving body image 601 that has not been detected can be detected.

  If all the moving body images 601 in the captured image cannot be properly detected, the background portion cannot be specified accurately. For this reason, when instructing redetection by the second redetection method, the operator is allowed to designate the background portion. Therefore, in the second re-detection method, the vicinity of the moving body image 601 that has not been detected is designated. The reason for specifying the neighborhood is that it is highly likely that the background model can be changed more appropriately. The designation is performed by a click operation by moving the mouse cursor 602 to the corresponding position, as in the first redetection method.

  The detection control unit 107 notified from the input unit 106 that such a click operation has been performed notifies the first moving body detection unit 103 of the position where the click operation has been performed, and performs re-detection. Instruct. At this time, the automatic adjustment unit 111 compares the pixel value between the background model and the photographed image only in the position and the vicinity thereof, and if there is a difference between the pixel value more than a predetermined value, The pixel value of the background model is changed according to the difference. Thus, the first moving body detection unit 103 detects the moving body image 601 after the automatic adjustment unit 111 changes the background model as necessary.

  The re-detection by the first moving body detection unit 103 is performed by the first or second re-detection method according to the re-detection setting content. Next, re-detection by the second moving body detection unit 104 will be described in detail. The re-detection is performed by any one of the third to fifth re-detection methods.

  In the third redetection method, the moving body image 601 is detected for a color image expressed by hue information, saturation information, and brightness information as a captured image. The detection is performed by combining the brightness information of one of the photographed image and the background model with the brightness information of the other. The change for matching the lightness information is performed by changing the lightness information of the background model to that of the photographed image, for example, as in the second redetection method.

  By changing the brightness information in such a manner, it is possible to absorb the difference in brightness due to the moving object, that is, to avoid the influence thereof. Thereby, it is possible to detect the moving object image 601 in which the influence of the difference in brightness of the moving object is reduced from the remaining hue information and saturation information. In order to change the brightness information, the operator performs the same designation as in the second redetection method. The detection control unit 107 notifies the second moving body detection unit 104 of the position information of the click operation input from the input unit 106 and instructs execution of re-detection.

FIG. 9 is a diagram for explaining a fourth redetection method performed by the second moving body detection unit 104.
In this re-detection method, the moving body image 601 is detected for a color image represented by each color information of red, green, and blue (RGB) as a captured image. The detection is performed assuming a coordinate system having each color information of RGB as coordinate axes. As the origin of the coordinate system, for example, perfect black or white is assumed. Thereby, for each pixel, each color information with respect to the origin is calculated, the direction of the straight line (color angle) connecting the coordinate position specified by the calculated color information and the origin, and the distance of the straight line (hereinafter referred to as “color distance”). )). The three arrows a to c with the origin as the base point shown in FIG. 9 each represent a straight line connecting the coordinate position of the different color and the origin. For this reason, a symbol corresponding to the color represented by the arrows is attached and described.

  The color a and the color b are in a relatively small range even if the color angles are substantially the same and the color distance is different. This happens when the same color is different to some extent in brightness. From this, the color a and the color b are actually regarded as the same or close colors. For the colors a and c, the color distances are almost the same, but the color angles are greatly different. For this reason, they are considered different colors. As such, if the color angle difference is within the allowable range, it is considered that there is no difference between the two colors, and if the difference is outside the allowable range, the difference outside the allowable range is further increased between the color distances. If so, it is considered that there is a difference between the two colors.

  For this reason, the moving body image 601 is detected by comparing the coordinate position between the background model and the captured image. Thereby, paying attention to the difference in color angle between the background model and the captured image, and then detecting the color difference between them in the form paying attention to the difference in color distance, the part to be regarded as the moving object image 601 is specified. Like to do.

  The moving object may be illuminated with light other than white light. In particular, under dark conditions, the shooting state of the moving object is greatly affected by the color of the illuminated light and the amount of light. For this reason, in the present embodiment, color information to be deleted can be specified among the three pieces of color information. Thereby, for example, in a situation where red illumination is present near the moving body, red information can be deleted or the amount of information can be reduced. In this way, more appropriate mobile object information 601 can be detected.

  The color information is specified by allowing the operator to select a color whose influence is desired to be eliminated. Unlike the first to third redetection methods, redetection by the fourth redetection method is performed by an operator instructing redetection. The detection control unit 107 notified from the input unit 106 that the instruction has been given instructs the second moving body detection unit 104 to perform re-detection by the fourth re-detection method.

  The photographed image is a still image or a moving image. When the captured image is a moving image, the detection control unit 107 detects the moving body image 601 retroactively from the time when the moving body image 601 is newly detected by an instruction of re-detection by the operator. The detection is performed using a redetection method in which the moving body image 601 is detected. Thereby, the moving image 601 is detected in the moving image for a period during which it can be detected.

  In this way, in this embodiment, when an operator designates a redetection method used for redetection in advance, and the operator instructs redetection, the moving body image 601 is redetected by the designated redetection method. It has become. The redetection method is different from the previously used detection method. Therefore, the detection condition of the moving body image 601 is automatically changed, and the moving body image 601 that could not be detected can be detected by changing the detection condition. Since the detection condition is automatically changed, more moving body images 601 can be detected more easily for the operator. For these reasons, high convenience is obtained for the operator.

  FIG. 2 is a flowchart of the moving object detection process. The moving body detection process is a process executed when a still image is acquired as a captured image. Next, the detection process will be described in detail with reference to FIG. The detection process is realized by causing the CPU mounted on the PC to execute the program as described above.

  First, in step S1, a range (length and area) and shape of the moving body image 601 at each position on the captured image are set. In subsequent step S <b> 2, a captured image is read from the image storage unit 102, and automatic measurement for detecting the moving body image 601 is performed. Thereafter, the process proceeds to step S3, and the detection result of the moving object image 601 is displayed on the display device 120. By executing the processing in step S2, the first moving body detection unit 103 is realized.

  In step S4 following step S3, it is determined whether or not a re-detection instruction has been issued. If the operator has received an instruction from the input unit 106 that the instruction has been given, the determination is yes and the process proceeds to step S5, where re-detection processing is performed to re-detect the mobile object image 601 according to the detection setting contents when preset. Execute. Thereafter, the process returns to step S3. On the other hand, if not, the determination is no and the process returns to step S3. In this manner, a state in which an instruction for redetection can be given is maintained until the operator gives an instruction to end.

  Redetection by the first to fourth redetection methods is performed by the redetection process according to the redetection setting content. Hereinafter, the operation for realizing the re-detection will be described in detail with reference to various processes shown in FIGS.

  FIG. 3 is a flowchart of the first detection process. The detection process is a process executed to realize redetection by the first redetection method. Next, the detection process will be described in detail with reference to FIG.

  First, in step S11, the position where the operator clicks from the input unit 106 is designated as a detection error point. In the next step S12, the previous detection level is set to an initial value. In the next step S13, the current detection level is set to ½ of the initial value. Thereafter, the process proceeds to step S14.

  In the background subtraction method, noise is caused by a slight change in the shooting location or a change in the background object (swing of a tree or the like). Therefore, in step S14, it is determined whether or not the current detection level is a noise region. If the detection level is the noise region, the determination is YES, the process proceeds to step S15, and a message indicating that the mobile object image 601 cannot be detected at the detection error point is displayed on the display device 120. Exit. Otherwise, the determination is no and the process moves to step S16. In step S14, it is also determined whether or not the absolute value of the difference between the current detection level and the detection level set in the past is equal to or less than a predetermined value. Thereby, when the absolute value of the difference becomes equal to or less than a predetermined value, it is determined as YES, and re-detection is avoided by changing the detection level more finely than necessary.

  In step S16, the moving body image 601 is detected at the current detection level. In the next step S17, it is determined whether or not the moving object image 601 has been detected by the detection. If the moving body image 601 cannot be detected, the determination is no and the process proceeds to step S18, where the current detection level and the center value of the previous detection level are set as the new current detection level. After the current detection level is set as the detection level, the process proceeds to step S14. Otherwise, the determination is no and the process moves to step S19.

  In step S19, the size of the detected moving body image 601 is compared with the reference, that is, the size range set in step S1. If the size is out of the larger range, it is determined and the process proceeds to step S18. If the size is out of the smaller range, the determination is made and the process proceeds to step S20, where the current detection level is set as the previous detection level and the current detection level is set as the new detection level. After setting 1/2, the process returns to step S14. If the size is within the range, it is determined that the moving body image 601 has been newly detected is displayed on the display device 120, and then the series of processing ends.

  FIG. 4 is a flowchart of the second detection process. The detection process is a process executed to realize redetection by the second redetection method. Next, the detection process will be described in detail with reference to FIG.

  First, in step S31, the position clicked by the operator from the input unit 106 is designated as the background portion. In the next step S32, an accumulated value or an average value of pixel value differences is calculated in correspondence between the designated background portion and corresponding portions in the applied background model. After the calculation, the process proceeds to step S33.

  In step S33, if the calculated accumulated value or average value is not within a predetermined allowable range, the pixel value of the background model is changed (adjusted) to match that of the photographed image. In the next step S34, the moving body image 601 is detected using the background model after the change. After performing the detection, the detection result is displayed on the display device 120, and then the series of processes is terminated.

  In the third re-detection method, after specifying the background portion in step S31, the brightness information of the background model is changed to match that of the photographed image, and the moving body image 601 is detected. Thereby, similarly to the second re-detection method, the influence of the difference in brightness due to the moving body on the detection of the moving body image 601 is avoided or reduced.

  FIG. 5 is a flowchart of the fourth detection process. The detection process is a process executed to realize redetection by the fourth redetection method. Finally, the detection process will be described in detail with reference to FIG. Here, only the processing executed after calculating the color angle and the color distance is shown.

  First, in step S41, the color angle difference between the background model and the captured image is determined. If there is a pixel whose difference (absolute value) is not within the allowable range (a pixel group that only satisfies the size range set in step S1 in FIG. 2), this is determined and the process proceeds to step S43. If not, that is, if the difference is within the allowable range, the fact is determined and the process proceeds to step S42, and the newly detected mobile object image 601 is shown on the display device 120. Thereafter, the series of processing is terminated.

  In step S43, a difference in color distance between them is determined. If the difference (absolute value) is within the allowable range, it is determined that the process proceeds to step S44, and the display device 120 displays that the mobile object image 601 could not be newly detected. Terminate the process. If not, that is, if the difference is outside the distance range, it is determined to that effect and the process proceeds to step S42.

  In the present embodiment, the detection of the moving body image 601 is first performed on a black and white image, but the detection may be performed on a color image. good. In that case, detection of the moving body image 601 using the background model before operating the brightness information may be performed first, and the re-detection may be performed using the background model after operating the brightness information. The re-detection may be performed by a fourth re-detection method. When a moving image is targeted as a captured image, an inter-frame difference method may be used. As apparent from the above, the detection condition changing method can be variously modified.

  The moving body detection apparatus according to the present embodiment is realized by a single PC. The moving body detection apparatus may be realized by a plurality of PCs.

It is a functional block diagram of the moving body detection apparatus by this Embodiment. It is a flowchart of a mobile body detection process. It is a flowchart of a 1st detection process. It is a flowchart of a 2nd detection process. It is a flowchart of a 4th detection process. It is a figure for demonstrating the display method of the detection result of a moving body image. It is a figure explaining the 1st re-detection method performed by the 1st mobile body detection part 103. FIG. It is a figure explaining the 2nd re-detection method performed by the 1st mobile body detection part 103. FIG. It is a figure explaining the 4th re-detection method performed by the 2nd mobile body detection part. It is a figure explaining the mobile body detection by a background difference method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Image acquisition part 102 Image memory | storage part 103 1st moving body detection part 104 2nd moving body detection part 105 Display control part 106 Input part 107 Detection control part 111 Automatic adjustment part 120 Display apparatus

Claims (13)

In a moving body detection apparatus that detects a moving body photographed as a subject in a captured image,
Image acquisition means for acquiring the captured image;
Moving body detection means for detecting a moving body image present in the captured image acquired by the image acquisition means;
Display control means for displaying a detection result by the moving body detection means together with the captured image on a display device;
Instruction means for instructing execution of re-detection of the moving object image;
Detection control means for changing the detection condition of the moving object image and causing the moving object detection means to execute detection of the moving object image again when execution of the re-detection is instructed by the instruction means;
A moving body detection apparatus comprising: The moving body detection means detects the moving body image from a difference between the captured image and a background image prepared in advance,
The detection control means changes the detection condition by changing a difference condition regarded as the moving body image.
The moving body detection apparatus according to claim 1. The re-detection instruction by the instruction means is performed by designating a moving body image displayed by the display control means and not detected by the moving body detection means,
The detection control means changes the condition of the difference applied to the position designated by the instruction means and the vicinity thereof,
The moving body detection apparatus according to claim 2. The moving body detection means detects the moving body image from a difference between the captured image and a background image prepared in advance,
The re-detection instruction by the instruction unit is performed by causing the display control unit to designate a portion where the moving body image is not displayed,
When the part is designated by the instruction means, the detection control means compares the designated part with the background image, and temporarily changes the background image based on the comparison result. Change detection conditions,
The moving body detection apparatus according to claim 1. The moving body detecting means includes first and second moving body detecting means that target black and white images and color images as the captured images, respectively.
The detection control means switches the object for detecting the moving body image from the first moving body detection means to the second moving body detection means when the instruction means instructs execution of the re-detection. To change the detection condition,
The moving body detection apparatus according to claim 1. The second moving body detection means treats a color image represented by hue information, saturation information, and brightness information as the photographed image, so that the difference between the color image and a background image prepared in advance is used. Detect moving body images,
The detection control means changes the lightness information of one of the color image and the background image to the lightness information of the other when the instruction means instructs execution of the re-detection, and Causing the moving body detection means to re-detect the moving body image;
The moving body detection apparatus according to claim 5. The second moving body detection means handles a color image represented by three color information of red, green, and blue as the color image, and assumes a coordinate system having the three color information as coordinate axes. Detecting the moving body image based on the position of the coordinate system represented by the three color information of the color image;
The moving body detection apparatus according to claim 5. The second moving body detecting means specifies a direction of a straight line connecting the origin of the coordinate system and the position, and a distance of the straight line, and detects the moving body image from a difference between the specified direction and the distance. I do,
The moving body detection device according to claim 7. The re-detection instruction by the instruction means can be performed by designating one of the three color information,
The detection control means operates the color information designated by the instruction means to cause the second moving body detection means to detect the moving body image;
The moving body detection device according to claim 7. The image acquisition means acquires a moving image as the captured image,
When the instruction means instructs the execution of the re-detection, the detection control means changes the detection condition and displays the moving image on the moving image displayed by the display control means when the instruction is given. The moving body detection means to detect the moving body image retroactively from the display position of
The moving body detection apparatus according to claim 1. In a moving body detection apparatus that detects a moving body photographed as a subject in a captured image,
Image acquisition means for acquiring a color image represented by three color information of red, green, and blue as the captured image;
Assuming a coordinate system having the three color information as coordinate axes, the position of the coordinate system represented by the three color information included in the color image acquired by the image acquisition unit is specified, and the specified position A moving body detection means for further specifying the direction of the straight line connecting the origin of the coordinate system and the distance of the straight line, and detecting the moving body image from the difference between the specified direction and the distance;
A moving body detection apparatus comprising: In a moving object detection method for detecting a moving object photographed as a subject in a captured image,
Detecting a moving body image present in the captured image, and displaying the detection result;
Re-detection for a moving body image that is recognized not to be detected by display of the detection result is performed by automatically changing detection conditions according to the re-detection instruction;
A moving body detection method characterized by the above. In a program to be executed by a computer used as a moving body detection device that detects a moving body photographed as a subject in a captured image,
An image acquisition function for acquiring the captured image;
A moving body detection function for detecting a moving body image present in a captured image acquired by the image acquisition function;
A display control function for displaying a detection result by the moving body detection function together with the captured image on a display device;
An instruction function for instructing execution of re-detection of the moving object image;
A detection control function for changing the detection condition of the moving body image and executing the detection of the moving body image again by the moving body detection function when execution of the re-detection is instructed by the instruction function;
A program to realize

Images