JP2006338611A - Moving object detection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a moving object detection device that can increase the accuracy of determining whether or not a detected moving object is present in a detection area. <P>SOLUTION: The moving object detection device 101, which processes image data acquired via an imaging device 102 to detect a moving object, comprises a lower region representative reference coordinate setting part 105 for setting lower region representative reference coordinates corresponding to a point representative of a lower region of a moving object area, and a detection area determination part 106 for determining whether or not the lower region representative reference coordinates are present in a predetermined detection area. The determination of whether or not the lower region representative reference coordinates corresponding to the point representative of the lower region of the moving object area are present in the predetermined detection area determines whether or not the moving object is present in the predetermined detection area. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a moving object detection apparatus with improved accuracy in determining whether or not a detected moving object exists within a detection region.

  2. Description of the Related Art Conventionally, a moving object detection and tracking processing method for detecting a moving object such as a human face area and tracking the movement of the moving object is known. In this moving object detection and tracking processing method, a human face area is cut out in a rectangular area having a preset size, and a labeling process is performed in the cut out rectangular area to obtain a centroid point of the human face area. Then, the movement of the detected moving object is stably tracked by obtaining the center point of the next cutout process based on the barycentric point (see, for example, Patent Document 1).

Further, in connection with the above-described conventional moving object detection and tracking processing method, a moving object detection device that notifies or reports an alarm when a moving object such as a person or an automobile enters a predetermined detection area is used. ing. In such a moving object detection device, the moving object region is extracted from the image data obtained from the imaging device, and the reference coordinates set as the coordinates of the center of gravity of the rectangular frame circumscribing the moving object region (that is, the coordinates of the center) are It is determined whether or not it exists in a predetermined detection area, and whether or not the feature amount of the moving object region satisfies a predetermined feature amount condition corresponding to a detection target (for example, a person), and the determination When it is determined that the detection target has entered the predetermined detection area based on the result, alarm notification or notification is performed.
Japanese Patent Laid-Open No. 5-68245 (page 4-7, FIG. 1)

  However, in the above moving object detection device, the coordinates of the center of the rectangular frame circumscribing the moving object area are set as the reference coordinates of the moving object area to determine whether or not the moving object exists in the detection area. Therefore, when a moving object moves near the boundary line of a predetermined detection area along the boundary line, the boundary line of the detection area is set obliquely, and the moving object becomes an oblique boundary line. When moving along, it is erroneously determined that a moving object that actually exists outside the detection area exists within the detection area, or a moving object that actually exists within the detection area is outside the detection area. If it exists, it is erroneously determined, and there is a problem that accuracy in determining whether or not the detected moving object exists in the detection region is low.

  The present invention has been made to solve the above-described conventional problems, and provides a moving object detection apparatus capable of improving the determination accuracy of whether or not a detected moving object exists in a detection region. With the goal.

  The moving object detection device of the present invention is a moving object detection device that detects image of a moving object by processing image data obtained by using an imaging device, the image data storing means for storing the image data, and the image data A moving object region extracting means for extracting a moving object region from the lower representative reference coordinate setting means for setting a lower representative reference coordinate corresponding to a point representing the lower portion of the moving object region, and the lower representative reference coordinate is a predetermined value. A detection area determination means for determining whether or not the detection object exists in the detection area; and a feature quantity for determining whether or not the feature quantity of the moving object area satisfies a predetermined feature quantity condition corresponding to the detection target moving object. A determination unit, and a notification unit that notifies that a detection target moving object is detected in the detection region based on the determination result of the detection region determination unit and the feature amount determination unit. The it has.

  With this configuration, whether or not the moving object exists in the predetermined detection area is determined by determining whether or not the lower representative reference coordinate corresponding to the point representing the lower part of the moving object area exists in the predetermined detection area. It can be determined whether or not. Therefore, when a moving object moves near the boundary of a predetermined detection area along the boundary, no matter how the boundary of the detection area is set, it actually exists outside the detection area. It is possible to reduce the probability that it is erroneously determined that the moving object exists within the detection area, or that the moving object that actually exists within the detection area is erroneously determined to exist outside the detection area.

  Further, in the moving object detection device of the present invention, the lower representative reference coordinate setting means is configured to set the lowest lower end of the moving object region as the lower representative reference coordinate of the moving object region.

  With this configuration, whether or not the moving object exists in the predetermined detection region is determined by determining whether or not the lowermost end of the moving object region set as the lower representative reference coordinate exists in the predetermined detection region. Can be determined. Therefore, when a moving object moves near the boundary of a predetermined detection area along the boundary, no matter how the boundary of the detection area is set, it actually exists in the detection area. The probability of erroneously determining that a moving object (for example, a moving object such as a person) exists outside the detection area can be reduced.

  Further, in the moving object detection device of the present invention, the lower representative reference coordinate setting means, when the rectangular frame circumscribing the moving object region is vertically long, uses the middle point of the bottom of the rectangular frame as the moving object region. When the rectangular frame circumscribing the moving object area is a horizontally long shape, the lower end of the moving object area is set as the lower representative reference coordinates of the moving object area. ing.

  With this configuration, when the rectangular frame circumscribing the moving object region is vertically long, for example, when the moving object is a person or the like, the midpoint of the bottom of the rectangular frame set as the lower representative reference coordinate is a predetermined detection. By determining whether or not the moving object exists in the area, it can be determined whether or not the moving object exists in the predetermined detection area. In addition, when the rectangular frame circumscribing the moving object area is horizontally long, for example, when the moving object is an automobile or the like, the lowermost end of the moving object area set as the lower representative reference coordinates exists within the predetermined detection area. By determining whether or not to perform, it is possible to determine whether or not the moving object exists within a predetermined detection area. Therefore, when a moving object moves near the boundary of a predetermined detection area along the boundary, no matter how the boundary of the detection area is set, and what kind of moving object Even if it is a moving object, it is erroneously determined that a moving object (for example, a moving body such as an automobile) that actually exists outside the detection area exists within the detection area, or actually exists within the detection area. The probability of erroneously determining that a moving object (such as a moving object such as a person) to be present outside the detection area can be reduced.

  Further, in the moving object detection device of the present invention, the lower representative reference coordinate setting means has a predetermined ratio of a portion existing in the detection area of the bottom to the entire bottom of the rectangular frame circumscribing the moving object area. The lower end of the moving object area is set as a lower representative reference coordinate of the moving object area, and the bottom edge of the rectangular frame circumscribing the moving object area is detected among the bottom edges. When the ratio of the portion existing in the area is larger than a predetermined ratio, the midpoint of the base of the rectangular frame is set to the lower representative reference coordinates of the moving object area.

  With this configuration, when the ratio of the portion existing in the detection area to the entire bottom of the rectangular frame circumscribing the moving object area is equal to or less than a predetermined ratio, for example, when the moving object is an automobile or the like. Determines whether the moving object exists in the predetermined detection area by determining whether the lowermost end of the moving object area set as the lower representative reference coordinates exists in the predetermined detection area. can do. In addition, when the ratio of the portion existing in the detection area of the bottom to the entire bottom of the rectangular frame circumscribing the moving object area is larger than a predetermined ratio, for example, when the moving object is a person, It is determined whether or not the moving object exists in the predetermined detection area by determining whether or not the midpoint of the bottom of the rectangular frame set as the lower representative reference coordinate exists in the predetermined detection area. be able to. Therefore, when a moving object moves near the boundary of a predetermined detection area along the boundary, no matter how the boundary of the detection area is set, and what kind of moving object Even if it is a moving object, it is erroneously determined that a moving object (for example, a moving body such as an automobile) that actually exists outside the detection area exists within the detection area, or actually exists within the detection area. The probability of erroneously determining that a moving object (such as a moving object such as a person) to be present outside the detection area can be reduced.

  Further, in the moving object detection device of the present invention, the lower representative reference coordinate setting means calculates the center of gravity of the rectangular frame circumscribing the moving object region and the center of gravity of the lower rectangular frame circumscribing the lower half of the moving object region. An intersection point where the straight line connecting and the bottom side of the rectangular frame intersect is set as the lower representative reference coordinate of the moving object region.

  With this configuration, the intersection set as the lower representative reference coordinate, that is, a straight line connecting the center of gravity of the rectangular frame circumscribing the moving object region and the center of gravity of the lower rectangular frame circumscribing the lower half of the moving object region, It can be determined whether or not the moving object exists in the predetermined detection region by determining whether or not the intersection where the bottom of the rectangular frame intersects exists in the predetermined detection region. Therefore, when a moving object moves near the boundary of a predetermined detection area along the boundary, no matter how the boundary of the detection area is set, and what kind of moving object Even if it is a moving object, it is possible to reduce the probability of erroneously determining that a moving object that actually exists outside the detection area (for example, a plurality of continuous moving objects such as automobiles) exists within the detection area. .

  The moving object detection device of the present invention is a moving object detection device that detects image of a moving object by processing image data obtained by using an imaging device, the image data storing means for storing the image data, and the image data Based on a moving object area extracting means for extracting a moving object area from the moving object area and a rectangular frame circumscribing the moving object area, a moving object area existing inside the rectangular frame and a predetermined area inside the rectangular frame When the ratio of the moving object area existing in the detection area is equal to or greater than a predetermined ratio, the detection area determining means for determining that the moving object exists in the detection area and the feature amount of the moving object area are detected. Feature amount determination means for determining whether or not a predetermined feature amount condition corresponding to the target moving object is satisfied, and based on the determination results of the in-detection region determination means and the feature amount determination means, Animal body has a configuration in which a notification means for notifying that it has been detected in the detection region.

  With this configuration, when the ratio of the moving object area existing inside the rectangular frame and in the predetermined detection area to the moving object area existing in the entire interior of the rectangular frame is equal to or higher than the predetermined ratio, It can be determined that the moving object exists within the detection area. Therefore, when a moving object moves near the boundary of a predetermined detection area along the boundary, no matter how the boundary of the detection area is set, it actually exists in the detection area. The probability of erroneously determining that a moving object (for example, a moving object such as a person) exists outside the detection area can be reduced.

  Further, the moving object detection device of the present invention is a moving object detection device for detecting a moving object by processing image data obtained using an imaging device, the image data storage means for storing the image data, Moving object area extraction means for extracting a moving object area from image data, and a rectangular frame circumscribing the moving object area has a vertically long shape, and all of the moving object areas existing inside the rectangular frame are subjected to predetermined detection. A detection area determination unit that determines that the moving object exists outside the detection area when the movement object is outside the detection area; and a predetermined feature amount condition that the feature value of the moving object area corresponds to the detection target moving object. A feature amount determination unit that determines whether or not the detection target is satisfied, and notifies that the moving object of the detection target is detected in the detection region based on the determination results of the detection region determination unit and the feature amount determination unit It has a configuration in which a notification means.

  With this configuration, even when the rectangular frame is vertically long, the moving object is outside the detection area when all of the moving object area existing inside the rectangular frame is outside the predetermined detection area. Then it can be determined. Therefore, when a moving object moves near the boundary of a predetermined detection area along the boundary, no matter how the boundary of the detection area is set, it actually exists outside the detection area. The probability of erroneously determining that a moving object (for example, a moving object such as an automobile) exists in the detection region can be reduced.

  The present invention determines a lower representative reference coordinate setting means for setting a lower representative reference coordinate corresponding to a point representing the lower part of a moving object region, and whether or not the lower representative reference coordinate exists within a predetermined detection region. By providing the detection area determination means, it is possible to provide a moving object detection device having an effect of improving the determination accuracy of whether or not a detected moving object exists in the detection area. It is.

  Hereinafter, a moving object detection device according to an embodiment of the present invention will be described with reference to the drawings. In this embodiment, a moving object detection device that performs notification when a detection target moving object (such as a person) is detected in a detection area (such as a sidewalk area) will be described as an example.

(First embodiment)
A block diagram of the moving object detection apparatus 101 according to the first embodiment of the present invention is shown in FIG. For example, a video signal of about 5 frames per second is input from the imaging device 102 such as a monitoring camera to the moving object detection device 101 of the present embodiment, and one input is input from the imaging device 102 to the moving object detection device 101. Image processing is performed on the video signal using a computer or the like. In FIG. 1, a moving object detection apparatus 101 includes a conversion / separation unit (not shown) that separates a video signal from the imaging apparatus 102 into a luminance signal and a color difference signal by filtering and analog / digital conversion, and a luminance signal. An image data storage unit 103 that stores image data including color difference signals, a moving object region extraction unit 104 that extracts a moving object region M from the image data stored in the image data storage unit 103, and a lower portion of the moving object region M. A lower representative reference coordinate setting unit 105 that sets a lower representative reference coordinate P corresponding to a representative point, and an in-detection region determination unit 106 that determines whether or not the lower representative reference coordinate P exists in a predetermined detection region R. A feature amount determination unit 107 that determines whether or not the feature amount of the moving object region M satisfies a predetermined feature amount condition corresponding to the detection target moving object, and determination within the detection region 106 and a notification unit 109 for notifying an alarm display device 108 such as an alarm device that a detection target moving object has been continuously detected a predetermined number of times in the detection area based on the determination result of the feature amount determination unit 107. And.

  The image data storage unit 103 includes a first image memory (not shown) that stores image data of the current frame, and a second image memory (not shown) that stores image data one frame before or several frames before. It has.

  The moving object region extraction unit 104 calculates a motion vector from the image data stored in the first image memory and the second image memory of the image data storage unit 103, and generates a motion generation block in which motion is assumed to be generated by the motion vector. A moving object region M is extracted by performing a labeling process. As a method for calculating the motion vector, for example, a gradient method for calculating a temporal change in luminance from image data stored in the first image memory and the second image memory and calculating a motion vector for each pixel, For example, a block matching method is used in which the image data is divided into blocks of about 8 pixels × 8 pixels and a correlation vector is evaluated for each block within a peripheral search range to calculate a motion vector.

  The lower representative reference coordinate setting unit 105 uses the lowermost block of the moving object region M extracted as described above as the lower representative reference coordinate P of the moving object region M corresponding to the point representing the lower part of the moving object region M. (See FIG. 3).

  A predetermined detection area R (for example, a sidewalk area) is set in advance in the detection area determination unit 106 (see FIG. 3). The detection area determination unit 106 is configured to detect a moving object (for example, a person) as a detection target within a predetermined detection area R (for example, a sidewalk area).

  In the feature amount determination unit 107, a predetermined feature amount condition of a moving object (for example, a person) as a detection target is set in advance. As the predetermined feature amount, for example, an aspect ratio, an area ratio, a width, a moving amount, a moving speed, and the like are used, and a predetermined feature amount condition is set for each feature amount. Here, the aspect ratio refers to a ratio A / B between the number of vertical blocks A of the rectangular frame F circumscribing the moving object region M and the number of horizontal blocks B. As the feature amount condition, for example, A / B The condition> 1 is set. Further, the area ratio refers to the area S (= A × B) of the rectangular frame F circumscribing the detected moving object region M and the rectangular frame F in the case of assuming the detection target moving object (for example, a person). The ratio S / So with the ideal area So is referred to as a feature amount condition, for example, a condition of 0.5 <S / So <3 is set. Further, the width W refers to a value obtained by converting the number of blocks B beside the rectangular frame F circumscribing the extracted moving object region M into a distance in the actual space, and the feature amount condition is, for example, 0.3 m < The condition of W <2m (more desirably 0.5m <W <1m) is set. The moving amount D is the lower representative reference coordinate P of the moving object region M from the lower representative reference coordinate P of the moving object region M when the moving object is first detected. A value obtained by converting the amount of movement to the farthest point into a distance in the actual space is set, and as a feature amount condition, for example, a condition of D> 1 m is set. Furthermore, the moving speed V is the speed at which the lower representative reference coordinate P of the moving object region M has moved from one frame to the next frame (for example, 0.2 seconds) is converted into a speed in actual space. As a feature amount condition, for example, a condition of V <10 km / h is set.

  The notification unit 109 includes a detection frequency determination unit (not shown), and a predetermined detection frequency (for example, 8 times) is preset in the detection frequency determination unit. The notification unit 109 detects when a detection target moving object (for example, a person) is detected within a detection region (for example, a sidewalk region) continuously by the detection number determination unit for a predetermined number of detections (for example, eight). The alarm display device 108 is configured to be notified. Note that a predetermined detection frequency ratio (50%) is preset in the detection frequency determination unit, and the notification unit 109 has a predetermined frequency ratio (for example, 50%) out of the predetermined detection frequency (for example, 8 times). As described above, the detection number determination unit determines that the feature amount (for example, aspect ratio, area ratio, etc.) of the moving object region M satisfies the predetermined feature amount condition corresponding to the detection target moving object (for example, person). In such a case, the alarm display device 108 may be notified.

  The operation of the moving object detection apparatus 101 configured as described above will be described using the flowchart shown in FIG.

  When image data obtained using the imaging device 102 is received from the first image memory and the second image memory of the image data storage unit 103, first, a motion vector is calculated based on the image data, and the motion is determined by the motion vector. A moving object area extraction process (S101) is performed in which the moving object area M is extracted by performing a labeling process on the motion generation block that is assumed to have occurred.

  Next, the lower representative reference coordinate setting process for setting the lowermost block of the extracted moving object region M to the lower representative reference coordinate P of the moving object region M corresponding to the point representing the lower part of the moving object region M ( S102) is performed (see FIG. 3).

  Then, in-detection area determination processing (S103) for determining whether or not the lower representative reference coordinate P, that is, the lowermost block of the moving object area M exists within a predetermined detection area R (for example, a sidewalk area) is performed. . As a result of the determination process in the detection area (S103), if it is determined that the lowermost block of the moving object area M exists in the predetermined detection area R (for example, a sidewalk area), the following feature amount calculation process (S104) ). On the other hand, when it is determined that the lowermost block of the moving object area M does not exist within the predetermined detection area R (for example, a sidewalk area), the detection of the moving object ends.

  As a result of the determination process within the detection area (S103), if it is determined that the lowermost block of the moving object area M exists within a predetermined detection area R (for example, a sidewalk area), the feature amount of the moving object area M A feature amount calculation process (S104) for calculating (eg, aspect ratio, area ratio, width, movement amount, movement speed, etc.) is performed.

  Then, a feature amount determination process (S105) is performed for determining whether or not the feature amount of the moving object satisfies a predetermined feature amount condition corresponding to the detection target moving object (for example, a person). As a result of the feature amount determination process (S105), when it is determined that the feature amount of the moving object satisfies a predetermined feature amount corresponding to the detection target moving object (for example, a person), the next detection frequency determination process The process proceeds to (S106). On the other hand, when it is determined that the feature amount of the moving object does not satisfy the predetermined feature amount condition corresponding to the detection target moving object (for example, a person), the detection of the moving object ends.

  As a result of the feature amount determination processing (S105), when it is determined that the feature amount of the moving object satisfies a predetermined feature amount condition corresponding to the detection target moving object (for example, a person), the detection target moving object ( For example, a detection number determination process (S106) is performed to determine whether or not a person or the like has been continuously detected within a detection region (for example, a sidewalk region) for a predetermined number of detections (for example, eight times). As a result of the detection number determination process (S106), when it is determined that the continuous detection number is a predetermined detection number (for example, 8 times), the subsequent detection result notification process (S107) is followed. On the other hand, if it is determined that the number of continuous detections is less than a predetermined number of detections (for example, 8 times) as a result of the detection number determination process (S106), the number of continuous detections of the detected moving object region M is 1 Then, the image data of the next frame is received, and the process returns to the moving object region extraction process (S101) again to perform the process sequentially.

  As a result of the detection number determination process (S106), when it is determined that the continuous detection number is a predetermined detection number (for example, eight times), the detection target moving object (for example, a person) is within the detection region (for example, a sidewalk). Detection result notification processing (S107) for notifying the alarm display device 108 of the detection in the area etc. is performed, and the detection of the moving object is completed.

  According to the moving object detection apparatus 101 of the first embodiment of the invention as described above, the lower representative reference coordinate setting unit 105 that sets the lower representative reference coordinates P corresponding to the point representing the lower part of the moving object region M; A detection area determination unit 106 that determines whether or not the lower representative reference coordinate P exists within the predetermined detection area R, and is the lowermost block of the moving object area M set as the lower representative reference coordinate P Can be accurately determined whether or not the moving object exists within the predetermined detection region R. Therefore, when a moving object moves near the boundary of a predetermined detection region R (for example, a sidewalk region) along the boundary, no matter how the boundary of the detection region R is set, Can reduce the probability of erroneously determining that a moving object (such as a person) existing in the detection region R exists outside the detection region R.

  For example, when the boundary line of the detection region (for example, a sidewalk region) R is set obliquely, as shown in FIG. 3, the detection region R is set with the coordinates of the center of the rectangular frame F circumscribing the moving object region M as the reference coordinates. Even if a determination is made whether or not a moving object (for example, a person) that exists in the detection region R actually exists outside the detection region R, it may be erroneously determined. In the moving object detection device 101 according to the first embodiment, since the lowermost block of the moving object area M is used as the lower representative reference coordinate P, it is determined whether or not the detection area R is within the detection area R. If a moving object (for example, a person, etc.) existing inside is outside the detection region R, the probability that it is erroneously determined is low, and therefore whether or not the detected moving object exists within the detection region R is determined. Can improve accuracy That.

(Second Embodiment)
Next, the moving object detection apparatus 201 according to the second embodiment of the present invention will be described. Here, the description will focus on differences from the first embodiment.

  In the first embodiment, the lower representative reference coordinate setting unit 105 is configured to set the lowermost block of the moving object region M to the lower representative reference coordinate P (see FIG. 3). In the second embodiment, the lower representative reference coordinate setting unit 205 sets a rectangular frame F circumscribing the moving object region M. When the rectangular frame F circumscribing the moving object region M is vertically long, When the middle point block of the bottom side of the rectangular frame F is set as the lower representative reference coordinate P of the moving object region M (see FIG. 5A), and the rectangular frame F circumscribing the moving object region M is horizontally long, The lowermost block of the moving object area M is set as the lower representative reference coordinate P of the moving object area M (see FIG. 5B).

  The operation of the moving object detection apparatus 201 configured as described above will be described with reference to the flowchart shown in FIG.

  In the first embodiment, the lowermost block of the moving object region M extracted by the moving object region extraction process (S101) is set to the lower part of the moving object region M corresponding to the point representing the lower part of the moving object region M. While the lower representative reference coordinate setting process (S102) for setting the representative reference coordinate P is performed (see FIGS. 2 and 3), in the second embodiment, the lower representative reference coordinate setting process as described below is performed. (S202) is performed (see FIG. 4).

  That is, after the moving object area extraction process (S201) similar to that of the first embodiment is performed, the rectangular frame setting process (S211) for setting the rectangular frame F circumscribing the extracted moving object area M is performed. Then, a shape determination process (S212) of the rectangular frame F is performed to determine whether the previous rectangular frame F has a vertically long shape or a horizontally long shape. As a result of the shape determination process (S212) of the rectangular frame F, when it is determined that the previous rectangular frame F has a vertically long shape, the middle representative block of the bottom side of the rectangular frame F is used as the lower representative reference coordinates of the moving object region M. P is set, and the lower representative reference coordinate setting process (S202) is terminated (see FIG. 5A). On the other hand, as a result of the shape determination process (S212) of the rectangular frame F, when it is determined that the previous rectangular frame F has a horizontally long shape, the lowermost block of the moving object region M is set as the lowermost block of the moving object region M. The coordinate is set to P, and the lower representative reference coordinate setting process (S202) is terminated (see FIG. 5B).

  According to the moving object detection device 201 of the second embodiment of the invention as described above, when the rectangular frame F circumscribing the moving object region M is vertically long, the block at the midpoint of the bottom of the rectangular frame F is determined. When the lower representative reference coordinate P of the moving object area M is set and the rectangular frame F circumscribing the moving object area M is horizontally long, the lowermost block of the moving object area M is used as the lower representative block of the moving object area M. By providing the lower representative reference coordinate setting unit 205 set to the coordinate P, when the rectangular frame F circumscribing the moving object region M is vertically long, for example, when the moving object is a person, the lower representative reference coordinate P By determining whether or not a block at the midpoint of the bottom side of the rectangular frame F set as is within a predetermined detection region R (for example, a sidewalk region), a moving object (for example, a person) is detected in the predetermined detection region. Is in R Or it can be determined (see Figure 5 (a)). When the rectangular frame F circumscribing the moving object area M is horizontally long, for example, when the moving object is an automobile or the like, the lowermost block of the moving object area M set as the lower representative reference coordinate P is a predetermined block. It is possible to determine whether or not a moving object (for example, an automobile or the like) exists in a predetermined detection region R by determining whether or not it exists in the detection region R (for example, a sidewalk region) (see FIG. 5 (b)). Therefore, when the moving object moves near the boundary line of a predetermined detection area R (for example, a sidewalk area) along the boundary line, no matter how the boundary line of the detection area R is set, No matter what kind of moving object the moving object is, it is erroneously determined that a moving object that actually exists outside the detection region R (for example, a moving object such as an automobile) exists within the detection region R, or In fact, it is possible to reduce the probability of erroneously determining that a moving object (for example, a moving object such as a person) existing in the detection region R exists outside the detection region R.

  For example, when the boundary line of the detection area (for example, a sidewalk area) R is set obliquely, as shown in FIG. 5A, the coordinates of the center of the rectangular frame F circumscribing the moving object area M are used as the reference coordinates. When it is determined whether or not it is within the detection region R, it may be erroneously determined that a moving object (such as a person) that actually exists within the detection region R exists outside the detection region R. Even in this case, the moving object detection device 201 according to the second embodiment determines whether the middle point block of the bottom of the rectangular frame F is within the detection region R as the lower representative reference coordinate P. Is less likely to erroneously determine that a moving object (such as a person) existing in the detection region R exists outside the detection region R, and therefore whether the detected moving object exists in the detection region R. Improve accuracy of judgment It can be.

(Third embodiment)
Next, the moving object detection apparatus 301 according to the third embodiment of the present invention will be described. Here, the description will focus on differences from the first embodiment.

  In the first embodiment, the lower representative reference coordinate setting unit 105 is configured to set the lowermost block of the moving object region M to the lower representative reference coordinate P (see FIG. 3). In the third embodiment, a predetermined ratio (for example, 50%) is preset in the lower representative reference coordinate setting unit 305. Then, the lower representative reference coordinate setting unit 305 sets a rectangular frame F circumscribing the moving object region M, and exists in the detection region R of the bottom with respect to the entire bottom of the rectangular frame F circumscribing the moving object region M. When the ratio of the portion to be moved is equal to or less than a predetermined ratio (for example, 50%), the lowermost block of the moving object area M is set as the lower representative reference coordinate P of the moving object area M (see FIG. 7A). ) If the ratio of the portion existing in the detection area R to the entire bottom edge of the rectangular frame F circumscribing the moving object area M is larger than a predetermined ratio, the midpoint of the bottom edge of the rectangular frame F Is set to the lower representative reference coordinate P of the moving object region M (see FIG. 7B).

  The operation of the moving object detection apparatus 301 configured as described above will be described with reference to the flowchart shown in FIG.

  In the first embodiment, the lowermost block of the moving object region M extracted by the moving object region extraction process (S101) is set to the lower part of the moving object region M corresponding to the point representing the lower part of the moving object region M. While the lower representative reference coordinate setting process (S102) for setting the representative reference coordinate P is performed (see FIGS. 2 and 3), in the third embodiment, the lower representative reference coordinate setting process as described below is performed. (S302) is performed (see FIG. 6).

  That is, after the moving object area extraction process (S301) similar to that of the first embodiment is performed, the rectangular frame setting process (S311) for setting the rectangular frame F circumscribing the extracted moving object area M is performed. The detection area of the bottom of the rectangular frame F for determining whether or not the ratio of the portion existing in the detection area R of the bottom to the entire bottom of the rectangular frame F is a predetermined ratio (for example, 50%) or less. An internal ratio determination process (S321) is performed. As a result of the above-described detection area ratio determination process (S321), it is determined that the ratio of the portion existing in the detection area R out of the bottom to the entire bottom of the rectangular frame F is equal to or less than a predetermined ratio (for example, 50%). If so, the lowermost block of the moving object area M is set as the lower representative reference coordinate P of the moving object area M, and the lower representative reference coordinate setting process (S302) is completed (see FIG. 7A). ). On the other hand, as a result of the above-described detection area ratio determination process (S321), the ratio of the portion existing in the detection area R out of the bottom to the entire bottom of the rectangular frame F is larger than a predetermined ratio (for example, 50%). Is determined, the midpoint block of the bottom of the rectangular frame F is set as the lower representative reference coordinate P of the moving object region M, and the lower representative reference coordinate setting process (S302) is completed (FIG. 7 ( b)).

  According to the moving object detection device 301 of the third embodiment of the invention as described above, the ratio of the portion existing in the detection region R out of the bottom to the entire bottom of the rectangular frame F circumscribing the moving object region M. Is equal to or less than a predetermined ratio, the lowermost block of the moving object region M is set as the lower representative reference coordinate P of the moving object region M, and the entire bottom side of the rectangular frame F circumscribing the moving object region M is set. If the ratio of the portion existing in the detection area R of the bottom side is larger than a predetermined ratio, the lower block for setting the middle point block of the bottom side of the rectangular frame F to the lower representative reference coordinate P of the moving object area M By providing the representative reference coordinate setting unit 305, the proportion of the portion existing in the detection region R (for example, the sidewalk region) of the bottom relative to the entire bottom of the rectangular frame F circumscribing the moving object region M is equal to or less than a predetermined proportion. For example, move When the body is an automobile or the like, the moving object is determined by determining whether or not the lowermost block of the moving object area M set as the lower representative reference coordinate P exists in the predetermined detection area R. It can be determined whether or not it exists within a predetermined detection region R (see FIG. 7A). Further, when the ratio of the portion existing in the detection area R (for example, the sidewalk area) to the entire bottom side of the rectangular frame F circumscribing the moving object area M is larger than a predetermined ratio, for example, the moving object is a person Or the like, by determining whether or not the block of the midpoint of the bottom side of the rectangular frame F set as the lower representative reference coordinate P exists in the predetermined detection region R, the moving object is determined in advance. It can be determined whether or not it exists in the detection region R (see FIG. 7B). Therefore, when the moving object moves near the boundary line of a predetermined detection area R (for example, a sidewalk area) along the boundary line, no matter how the boundary line of the detection area R is set, No matter what kind of moving object the moving object is, it is erroneously determined that a moving object (for example, a moving body such as a car) that actually exists outside the detection region R exists within the detection region R, or In fact, it is possible to reduce the probability of erroneously determining that a moving object (for example, a moving body such as a person) existing in the detection region R exists outside the detection region R.

  For example, when the boundary line of the detection region (for example, the sidewalk region) R is set obliquely, as shown in FIG. 7B, the coordinates of the center of the rectangular frame F circumscribing the moving object region M are used as the reference coordinates. When it is determined whether or not it is within the detection region R, it may be erroneously determined that a moving object (such as a person) that actually exists within the detection region R exists outside the detection region R. Even in this case, the moving object detection device 301 according to the third embodiment determines whether or not the middle point block of the bottom of the rectangular frame F is within the detection region R as the lower representative reference coordinate P. Is less likely to erroneously determine that a moving object (such as a person) existing in the detection region R exists outside the detection region R, and therefore whether the detected moving object exists in the detection region R. Improve accuracy of judgment It can be.

(Fourth embodiment)
Next, a moving object detection device 401 according to a fourth embodiment of the present invention will be described. Here, the description will focus on differences from the first embodiment.

  In the first embodiment, the lower representative reference coordinate setting unit 105 is configured to set the lowermost block of the moving object region M to the lower representative reference coordinate P (see FIG. 3). In the fourth embodiment, the lower representative reference coordinate setting unit 405 sets a rectangular frame F circumscribing the moving object region M, and calculates the center of gravity G of the rectangular frame F circumscribing the moving object region M. Further, the lower representative reference coordinate setting unit 405 sets the lower rectangular frame f circumscribing the lower half of the moving object region M, and calculates the center of gravity g of the lower rectangular frame f circumscribing the lower half of the moving object region M. To do. Then, the lower representative reference coordinate setting unit 405 calculates an intersection where a straight line connecting the block of the center of gravity G of the rectangular frame F and the block of the center of gravity g of the lower rectangular frame f intersects the bottom of the rectangular frame F, The block of the intersection is set to the lower representative reference coordinate P of the moving object region M (see FIG. 9).

  The operation of the moving object detection apparatus 401 configured as described above will be described with reference to the flowchart shown in FIG.

  In the first embodiment, the lowermost block of the moving object region M extracted by the moving object region extraction process (S101) is set to the lower part of the moving object region M corresponding to the point representing the lower part of the moving object region M. While the lower representative reference coordinate setting process (S102) for setting the representative reference coordinates P is performed (see FIGS. 2 and 3), the lower representative reference coordinate setting process as described below is performed in the fourth embodiment. (S402) is performed (see FIG. 8).

  That is, after the moving object area extraction process (S401) similar to that of the first embodiment is performed, first, a rectangular frame setting process for setting a rectangular frame F circumscribing the extracted moving object area M is performed. Then, the center-of-gravity calculation processing (S431) of the rectangular frame F for calculating the block of the center of gravity G of the rectangular frame F is performed. Next, a lower rectangular frame setting process for setting a lower rectangular frame f circumscribing the lower half portion of the moving object region M is performed, and a gravity center calculation process for the lower rectangular frame f for calculating a block of the gravity center g of the lower rectangular frame f. (S433) is performed. Then, an intersection calculation process (S434) is performed to calculate a block of an intersection where a straight line connecting the block of the center of gravity G of the rectangular frame F and the block of the center of gravity g of the lower rectangular frame f intersects the bottom of the rectangular frame F. Is set as the lower representative reference coordinate P of the moving object region M, and the lower representative reference coordinate setting process (S402) is completed (see FIG. 8).

  According to the moving object detection device 401 of the fourth embodiment of the invention as described above, the block of the center of gravity G of the rectangular frame F circumscribing the moving object region M and the lower rectangle circumscribing the lower half of the moving object region M. A lower representative reference coordinate setting unit 405 is provided for setting a block at an intersection where a straight line connecting the block of the center of gravity g of the frame f and the bottom side of the rectangular frame F intersects with the lower representative reference coordinate P of the moving object region M. Thus, by determining whether or not the block of the intersection set as the lower representative reference coordinate P exists within the predetermined detection region R, the moving object (for example, a continuous moving body such as a plurality of automobiles) is predetermined. It can be determined whether or not it exists in the detection region R (for example, a sidewalk region). That is, when a moving object moves near a boundary line of a predetermined detection region R (for example, a sidewalk region) along the boundary line, no matter how the boundary line of the detection region R is set, for example, Even if the boundary line of the detection region R is set obliquely, and the moving object is any type of moving object (for example, a moving body such as a plurality of continuous cars), the moving object region M A straight line connecting the block of the center of gravity G of the rectangular frame F circumscribing to the block of the center of gravity g of the lower rectangular frame f circumscribing the lower half of the moving object region M is set obliquely along the boundary line. . Therefore, the probability of erroneously determining that a moving object that actually exists outside the detection region R exists within the detection region R can be reduced. Therefore, it is possible to improve the determination accuracy of whether or not the detected moving object exists in the detection region R.

(Fifth embodiment)
Next, a block diagram of the moving object detection device 501 according to the fifth embodiment of the present invention is shown in FIG. In FIG. 10, a moving object detection device 501 converts and separates a video signal from an imaging device 502 such as a monitoring camera into a luminance signal and a color difference signal by filtering and analog / digital conversion (not shown). An image data storage unit 503 that stores image data including a luminance signal and a color difference signal, a moving object region extraction unit 504 that extracts a moving object region M from the image data stored in the image data storage unit 503, and a moving object Based on a rectangular frame F circumscribing the area M, a detection area determination unit 506 that determines whether or not a moving object exists within a predetermined detection area R, and a feature amount of the moving object area M is a detection target moving object. Is detected based on the determination results of the feature amount determination unit 507 that determines whether or not a predetermined feature amount condition corresponding to the condition is satisfied, and the determination result in the detection region determination unit 506 and the feature amount determination unit 507 Moving object mark has a notification unit 509 to notify the alarm display device 508 such as that which is continuously detected a predetermined number of detection times in the detection region e.g. alarm device.

  A predetermined ratio (for example, 15%) is set in advance in the detection area determination unit 506. Then, the in-detection area determination unit 506 blocks the moving object area M existing inside the rectangular frame F and in the predetermined detection area R with respect to the number of blocks of the moving object area M existing inside the entire rectangular frame F. When the ratio of the number is a predetermined ratio (for example, 15%) or more, it is determined that the moving object exists in the detection region R (see FIG. 12). Note that the image data storage unit 503, the moving object region extraction unit 504, the feature amount determination unit 507, and the notification unit 509 have the same configurations as those in the first embodiment, and thus description thereof is omitted here.

  The operation of the moving object detection apparatus 501 configured as described above will be described with reference to the flowchart shown in FIG.

  When image data obtained using the imaging device 102 is received from the first image memory and the second image memory of the image data storage unit 103, first, a motion vector is calculated based on the image data, and the motion is determined by the motion vector. A moving object region extraction process (S501) is performed in which the moving object region M is extracted by applying a labeling process to the motion generation block that is assumed to have occurred.

  Next, a rectangular frame setting process (S511) for setting a rectangular frame F circumscribing the extracted moving object region M is performed, and a rectangular shape is set with respect to the number of blocks of the moving object region M existing inside the rectangular frame F. In-detection area determination processing for determining whether or not the ratio of the number of blocks in the moving object area M existing within the frame F and within the predetermined detection area R is equal to or greater than a predetermined ratio (for example, 15%) (S503). Is done. As a result of the detection area inside determination process (S503), the number of blocks of the moving object area M existing in the entire inside of the rectangular frame F is equal to that of the moving object area M existing in the rectangular frame F and in the predetermined detection area R. When it is determined that the ratio of the number of blocks is equal to or greater than a predetermined ratio (for example, 15%), the process proceeds to the next feature amount calculation process (S504) (see FIG. 12). On the other hand, as a result of the detection area inside determination process (S503), the moving object area existing inside the rectangular frame F and within the predetermined detection area R with respect to the number of blocks of the moving object area M existing inside the entire rectangular frame F. When it is determined that the ratio of the number of M blocks is smaller than a predetermined ratio (for example, 15%), the moving object detection process ends.

  Thereafter, a feature amount calculation process (S504), a feature amount determination process (S505), a detection count determination process (S506), and a detection result notification process (S507) are performed. These processes are described in the first embodiment. Since the same processing is performed, the description is omitted here.

  According to the moving object detection device 501 of the fifth embodiment of the invention as described above, the block of the moving object region M existing in the entire interior of the rectangular frame F based on the rectangular frame F circumscribing the moving object region M. If the ratio of the number of blocks in the moving object area M existing in the rectangular frame F and in the predetermined detection area R is greater than or equal to a predetermined ratio (for example, 15%), the moving object is detected in the detection area R. By providing the detection area determination unit 506 that determines that the object exists in the area, it is possible to accurately determine that the moving object (for example, a person) exists in the detection area R (for example, the sidewalk area) (see FIG. 12). Therefore, when a moving object (for example, a person) moves near the boundary of a predetermined detection region R (for example, a sidewalk region) along the boundary, how is the boundary of the detection region R set? However, the probability of erroneously determining that a moving object that actually exists in the detection region R exists outside the detection region R can be reduced.

  For example, when the boundary line of the detection region (for example, a sidewalk region) R is set obliquely, the coordinates of the center of the rectangular frame F circumscribing the moving object region M are detected based on the reference coordinates as shown in FIG. When it is determined whether or not it is within the region R, it may be erroneously determined that a moving object (such as a person) that exists in the detection region R actually exists outside the detection region R. However, in the moving object detection device 501 of the fifth embodiment, the movement existing inside the rectangular frame F and within the predetermined detection area R with respect to the number of blocks of the moving object area M existing inside the rectangular frame F. Since it is determined whether or not the detection area R is within the detection area R based on the ratio of the number of blocks in the object area M, it is erroneous if a moving object (such as a person) that actually exists in the detection area R exists outside the detection area R The probability of judging , Therefore, can be detected moving object is to improve the accuracy of determining whether present in the detection area R.

(Sixth embodiment)
Next, the moving object detection device 601 according to the sixth embodiment of the present invention will be described. Here, the description will focus on differences from the fifth embodiment.

  In the fifth embodiment, the in-detection area determination unit 506 exists inside the rectangular frame F and in the predetermined detection area R with respect to the number of blocks of the moving object area M existing in the entire inside of the rectangular frame F. When the ratio of the number of blocks in the moving object area M is equal to or greater than that ratio, the moving object area M is configured to determine that the moving object exists in the detection area R (see FIG. 11). In the embodiment, the in-detection area determination unit 606 has the rectangular frame F in a vertically long shape, and all the blocks of the moving object area M existing inside the rectangular frame F exist outside the predetermined detection area R. In such a case, it is determined that the moving object exists outside the detection region R (see FIG. 13).

  The operation of the moving object detection apparatus 601 configured as described above will be described using the flowchart shown in FIG.

  In the fifth embodiment, the inside of the rectangular frame F and within the predetermined detection region R with respect to the number of blocks of the moving object region M existing in the entire interior of the rectangular frame F set by the rectangular frame setting process (S511). In the sixth embodiment, the detection area determination process (S503) for determining whether or not the ratio of the number of blocks of the moving object area M existing in is equal to or greater than a predetermined ratio is performed in the sixth embodiment. The following processing is performed (see FIG. 13).

  That is, after the rectangular frame setting process (S611) similar to that of the fifth embodiment is performed, first, the shape of the rectangular frame F set by the rectangular frame setting process (S611) is a vertically long shape or a horizontally long shape. When the shape determination process (S612) of the rectangular frame F for determining whether the rectangular frame F is a vertically long shape is determined as a result of the shape determination process (S612), the following detection is performed. The process proceeds to the in-region determination process (S603) (see FIG. 14). On the other hand, as a result of the shape determination process (S612), when it is determined that the shape of the rectangular frame F is a horizontally long shape, the detection of the moving object ends.

  Then, as a result of the shape determination process (S612), when it is determined that the shape of the rectangular frame F is a vertically long shape, all the blocks of the moving object region M existing inside the rectangular frame F are predetermined detection regions. In-detection area determination processing (S603) for determining whether or not the object exists outside R is performed. As a result of the determination process within the detection area (S603), if it is determined that at least one block of the moving object area M existing inside the rectangular frame F exists outside the predetermined detection area R, the following feature amount calculation process is performed. The process proceeds to (S604). On the other hand, as a result of the determination process within the detection area (S603), if it is determined that all the blocks of the moving object area M existing inside the rectangular frame F exist outside the predetermined detection area R, the detection of the moving object is performed. Ends (see FIG. 14).

  According to the moving object detection device 601 of the sixth embodiment of the present invention, the rectangular frame F circumscribing the moving object region M has a vertically long shape, and the moving object region exists inside the rectangular frame F. When all the blocks of M exist outside the predetermined detection area R, the detection area inside determination unit 606 that determines that the moving object exists outside the detection area R is provided, so that the moving object (for example, an automobile) is present. It can be accurately determined that the object exists outside the detection region R (for example, a sidewalk region).

  For example, in the case of rainy weather, a vertically long rectangular frame F may be set for a moving object (for example, an automobile) to which a horizontally long rectangular frame F should be originally set. 14), even if the rectangular frame F has a vertically long shape, if the entire moving object region M existing inside the rectangular frame F exists outside the predetermined detection region R (for example, a sidewalk region), Can reduce the probability of erroneously determining that a moving object existing outside the detection region R exists within the detection region R. In addition, when a moving object (for example, an automobile) moves along the boundary line of a predetermined detection region R (for example, a sidewalk region), how is the boundary line of the detection region R set? For example, even if the boundary line of the detection region R is set obliquely, the probability of erroneously determining that a moving object that actually exists outside the detection region R exists in the detection region R can be reduced. it can. Therefore, it is possible to improve the determination accuracy of whether or not the detected moving object exists in the detection region R.

  The preferred embodiments of the present invention have been described above. However, the scope of the present invention is not limited to the above-described embodiments, and may be modified according to the purpose within the scope of the claims. It is possible.

  In the fourth embodiment, the lower representative reference coordinate setting unit 405 includes the block of the center of gravity G of the rectangular frame F circumscribing the moving object region M and the lower rectangular frame f circumscribing the lower half of the moving object region M. An example has been described in which an intersection where the straight line connecting the block of the center of gravity g and the bottom of the rectangular frame F intersect is calculated and the block at the intersection is set as the lower representative reference coordinate P of the moving object region M. Otherwise, as shown in FIG. 15, the lower representative reference coordinate setting unit 405 includes a block of the center of gravity G of the rectangular frame F circumscribing the moving object region M and a motion generation block region (motion of the lower half of the moving object region M). The intersection point between the straight line connecting the blocks of the center of gravity g of the generated block) and the bottom side of the rectangular frame F is calculated, and the block at the intersection point is set as the lower representative reference coordinate P of the moving object region M. Can be similarly implemented.

  By determining whether or not the intersection block set as the lower representative reference coordinate P exists in the predetermined detection area R as described above, a moving object (for example, a moving object such as a plurality of continuous cars) ) Exists in a predetermined detection region R (for example, a sidewalk region). That is, when a moving object moves near the boundary line of a predetermined detection area R (for example, a sidewalk area) along the boundary line, even if the boundary line of the detection area R is set obliquely, Regardless of the type of moving object (for example, a plurality of continuous moving objects such as automobiles), the block of the center of gravity G of the rectangular frame F circumscribing the moving object area M and the lower half of the moving object area M A straight line connecting the block of the center of gravity g of the motion generation block area (collection of motion generation blocks) of the part is set obliquely along the boundary line. For example, in another embodiment shown in FIG. 15, the center of gravity g of the motion generation block region in the lower half of the moving object region M is set on the right side as compared to the fourth embodiment shown in FIG. The As a result, the slope of the straight line connecting the centroid G and the centroid g becomes closer to the slope of the boundary line, and the intersection where the straight line and the bottom of the rectangular frame F intersect is set to the right. Therefore, the probability of erroneously determining that a moving object that actually exists outside the detection region R exists within the detection region R can be further reduced. Therefore, it is possible to further improve the determination accuracy of whether or not the detected moving object exists in the detection region R.

  As described above, the moving object detection device according to the present invention has an effect that it is possible to improve the determination accuracy of whether or not the detected moving object exists in the detection region, such as a person or a car. This is useful as a moving object detection device for notifying or reporting an alarm when a moving object enters a predetermined detection area.

The block diagram of the moving object detection apparatus in the 1st Embodiment of this invention The flowchart for demonstrating operation | movement of the moving object detection apparatus in the 1st Embodiment of this invention. Explanatory drawing of the lower representative reference coordinates in the first embodiment of the present invention Flowchart of lower representative reference coordinate setting process in the second embodiment of the present invention (A) Explanatory drawing of lower representative reference coordinates in the second embodiment of the present invention (when the rectangular frame is vertically long) (b) Second embodiment of the present invention (when the rectangular frame is horizontally long) Explanatory drawing of lower representative reference coordinates in Flowchart of lower representative reference coordinate setting process in the third embodiment of the present invention (A) Explanatory drawing of lower representative reference coordinates in the third embodiment of the present invention (when the ratio of the bottom side of the rectangular frame within the detection area is a predetermined ratio or less) (b) Third embodiment of the present invention ( Explanatory drawing of lower representative reference coordinates in the case where the ratio of the bottom of the rectangular frame in the detection area is greater than a predetermined ratio) Flowchart of lower representative reference coordinate setting process in the fourth embodiment of the present invention Explanatory drawing of the lower representative reference coordinates in the fourth embodiment of the present invention The block diagram of the moving object detection apparatus in the 5th Embodiment of this invention Flowchart for explaining the operation of the moving object detection device according to the fifth embodiment of the present invention. Explanatory drawing of the determination in a detection area | region in the 5th Embodiment of this invention Flowchart for explaining the operation of the moving object detection device according to the sixth embodiment of the present invention. Explanatory drawing of the determination in a detection area in the 6th Embodiment of this invention. Explanatory drawing of the lower representative reference coordinates in another embodiment of the present invention

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Moving object detection apparatus 102 Imaging apparatus 103 Image data memory | storage part 104 Moving object area | region extraction part 105 Lower representative reference coordinate setting part 106 Detection area inside determination part 107 Feature-value determination part 109 Notification part M Moving object area | region P Lower representative reference coordinate R Detection area F Rectangular frame

Claims (7)

A moving object detection device for detecting a moving object by processing image data obtained using an imaging device,
Image data storage means for storing the image data; moving object area extraction means for extracting a moving object area from the image data; and a lower part for setting lower representative reference coordinates corresponding to a point representing the lower part of the moving object area Representative reference coordinate setting means, detection area determination means for determining whether or not the lower representative reference coordinates exist within a predetermined detection area, and a feature amount of the moving object area corresponds to a moving object as a detection target. A feature amount determination unit that determines whether or not a predetermined feature amount condition is satisfied, and a detection target moving object is detected in the detection region based on the determination results of the detection area determination unit and the feature amount determination unit A moving object detection apparatus comprising: notification means for notifying the above.   The moving object detection device according to claim 1, wherein the lower representative reference coordinate setting unit sets a lowermost representative reference coordinate of the moving object area as a lowermost end of the moving object area.   When the rectangular frame circumscribing the moving object region is vertically long, the lower representative reference coordinate setting means sets the midpoint of the bottom of the rectangular frame as the lower representative reference coordinate of the moving object region, and the movement 2. The moving object detection device according to claim 1, wherein when the rectangular frame circumscribing the object area is horizontally long, the lowermost end of the moving object area is set as a lower representative reference coordinate of the moving object area. .   The lower representative reference coordinate setting means, when the ratio of the portion existing in the detection area of the bottom side to the entire bottom side of the rectangular frame circumscribing the moving object area is equal to or less than a predetermined ratio, The lowest end of the object area is set as the lower representative reference coordinate of the moving object area, and the ratio of the portion existing in the detection area of the bottom edge to the entire bottom edge of the rectangular frame circumscribing the moving object area is predetermined. 2. The moving object detection device according to claim 1, wherein when the ratio is larger than the ratio, the midpoint of the bottom of the rectangular frame is set as a lower representative reference coordinate of the moving object region.   The lower representative reference coordinate setting means includes a straight line connecting a centroid of a rectangular frame circumscribing the moving object region and a centroid of a lower rectangular frame circumscribing the lower half of the moving object region, and a bottom side of the rectangular frame. The moving object detection apparatus according to claim 1, wherein an intersecting intersection point is set to a lower representative reference coordinate of the moving object region. A moving object detection device for detecting a moving object by processing image data obtained using an imaging device,
Based on an image data storage means for storing the image data, a moving object area extraction means for extracting a moving object area from the image data, and a rectangular frame circumscribing the moving object area, the image data storage means exists in the entire interior of the rectangular frame. If the ratio of the moving object area existing inside the rectangular frame and in the predetermined detection area to the moving object area to be detected is equal to or greater than the predetermined ratio, the movement is determined to be present in the detection area. In-detection area determination means for determining whether or not an object exists in a predetermined detection area, and whether or not the feature quantity of the moving object area satisfies a predetermined feature quantity condition corresponding to the detection target moving object. A feature amount determination unit for determining, and a notification unit for notifying that the moving object of the detection target is detected in the detection region based on the determination result of the detection region determination unit and the feature amount determination unit. Moving object detection apparatus according to claim and. A moving object detection device for detecting a moving object by processing image data obtained using an imaging device,
Image data storage means for storing the image data, moving object area extraction means for extracting a moving object area from the image data, a rectangular frame circumscribing the moving object area is a vertically long shape, and the rectangular frame When all the moving object areas existing inside are outside the predetermined detection area, a detection area determination unit that determines that the moving object exists outside the detection area, and a feature amount of the moving object area are detected. A feature amount determination unit that determines whether or not a predetermined feature amount condition corresponding to the target moving object is satisfied, and a detection target moving object is determined based on the determination results of the in-detection region determination unit and the feature amount determination unit. A moving object detection device comprising: notification means for notifying that detection has been made within a detection area.

Images