JP3577875B2 - Moving object extraction device - Google Patents

Moving object extraction device Download PDF

Info

Publication number
JP3577875B2
JP3577875B2 JP05132397A JP5132397A JP3577875B2 JP 3577875 B2 JP3577875 B2 JP 3577875B2 JP 05132397 A JP05132397 A JP 05132397A JP 5132397 A JP5132397 A JP 5132397A JP 3577875 B2 JP3577875 B2 JP 3577875B2
Authority
JP
Japan
Prior art keywords
template
position
moving
camera
moving object
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
JP05132397A
Other languages
Japanese (ja)
Other versions
JPH10255057A (en
Inventor
基孫 中
和史 水澤
武久 田中
利和 藤岡
Original Assignee
松下電器産業株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 松下電器産業株式会社 filed Critical 松下電器産業株式会社
Priority to JP05132397A priority Critical patent/JP3577875B2/en
Publication of JPH10255057A publication Critical patent/JPH10255057A/en
Application granted granted Critical
Publication of JP3577875B2 publication Critical patent/JP3577875B2/en
Anticipated expiration legal-status Critical
Application status is Expired - Fee Related legal-status Critical

Links

Images

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a technique for detecting a moving area from a moving image and accurately extracting a moving object in a monitoring device or the like.
[0002]
[Prior art]
In recent years, monitoring systems using a large number of cameras have been increasing for wide area monitoring and facility monitoring. For this reason, the development of an automatic monitoring system capable of grasping the situation by capturing the movement of a target object, etc. is expected, excluding an inefficient monitoring work in which an observer monitors a plurality of videos. In particular, as a method of tracking a moving object in a wide observation area with multiple camera images, an area of the target object in the image is accurately extracted, a foot position of the target object is obtained, and the position is determined in the observation space. To track the target object, or image the same observation area with two cameras and associate the same object in the two images, and calculate the parallax of the two cameras. There is a method of performing stereo processing for obtaining a depth distance of a target object using the method.
[0003]
[Problems to be solved by the invention]
However, the conventional technique has a major problem. For example, in the method of calculating the foot position of the target object, it is difficult to obtain the foot position when the overlap of the target object, the position of the stationary object, and the lower part of the target are hidden by a shield. Also, in the stereo processing, the same pattern matching in the video has difficulty such as the presence of a target object having little edge information and a target object whose pattern cannot be matched due to parallax.
[0004]
SUMMARY OF THE INVENTION The present invention has been made in view of the above-described conventional problems, and has been made accurate by introducing a multi-viewpoint process from a template position based on an automatically generated template matching method with a small processing load as a method of integrating a plurality of videos. The purpose is to extract a moving object.
[0005]
[Means for Solving the Problems]
In order to solve the problem, the present invention provides input means for inputting a camera image, moving area extracting means for detecting a moving area from video data obtained from the input means, and moving area extracting means. Area for each moving area from the moving areaSize and number of pixelsLabeling processing means for detecting information on a moving area, and using the information on each moving area detected by the labeling processing means,WhenTarget moving objectSimilarity withShape information extracting means for extracting,Based on the similarityVideo data obtained from the input means and information on a moving area obtained from the shape information extracting meansWhen the template already exists in the search range, the closest template is updated, and when the template does not exist in the range, the template is generated, and the information of the template is output as the information of the moving object.A camera image processing system having a template processing unit, an integrated processing unit for integrating information from the plurality of camera image processing systems, and processing of the position and classification information of a moving object obtained from the integrated processing unit.outputOutput processing means.
[0006]
Accordingly, in the present invention, a plurality of camera images having a known mapping function between a spatial position and a pixel position in each image are used in a processing procedure that makes use of the feature of the moving image processing to take advantage of the moving image processing. By extracting foot position information and performing multi-viewpoint processing and time-series processing on this unstable information, it is possible to obtain a highly accurate motion trajectory (trace) of the target area and accurately extract a moving object. Can be.
[0007]
BEST MODE FOR CARRYING OUT THE INVENTION
According to a first aspect of the present invention, there is provided an input unit for inputting a camera image, a moving region extracting unit for detecting a moving region from video data obtained from the input unit, and a moving region extracting unit. Regions for each moving region from the obtained moving regionSize and number of pixelsLabeling processing means for detecting information on a moving area, and using the information on each moving area detected by the labeling processing means,WhenTarget moving objectSimilarity withShape information extracting means for extracting,Based on the similarityVideo data obtained from the input means and information on a moving area obtained from the shape information extracting meansWhen the template already exists in the search range, the closest template is updated, and when the template does not exist in the range, the template is generated, and the information of the template is output as the information of the moving object.A camera image processing system having a template processing unit, an integrated processing unit for integrating information from the plurality of camera image processing systems, and processing of the position and classification information of a moving object obtained from the integrated processing unit.outputThis is a moving object extraction device composed of output processing means that performs multi-viewpoint moving object information using template processing combining shape feature extraction means and template matching method for tracking a moving object. The integration / identification processing has an effect that it is possible to track a moving object with high accuracy by simple processing utilizing characteristics of moving image processing.
[0008]
The invention according to claim 2 of the present invention is characterized in that, in each of the observation areas photographed by a plurality of cameras, each of the observation areas is photographed in duplicate by two or more camera images. This is a moving object extracting device, and has an effect that the effect of the invention described in claim 1 can be further enhanced.
[0009]
The invention described in claim 3 of the present invention provides a camera 1, a camera 2, a camera 3, and a camera 4 when four cameras are set as a set when a plurality of cameras are installed. Cameras 1 and 2 and cameras 3 and 4 are installed facing each other in the direction where cameras 2 and 3 and 4 face in parallel or inward, and any observation area can be photographed with two or more camera images in duplicate. The moving object extraction device according to the first or second aspect is characterized in that the number of cameras is increased as described above, and the effects of the first and second aspects can be obtained in a long stretched area such as a highway. Has an action.
[0010]
According to a fourth aspect of the present invention, there is provided the moving object extracting apparatus according to any one of the first to third aspects, wherein the input means is a visible camera, an infrared camera, or a combination thereof. This is a limitation of the input means for achieving the effects of the inventions described in 1 to 3, and has an effect that the advantages can be obtained by compensating for the advantages and disadvantages of both the visible camera and the infrared camera.
[0011]
According to a fifth aspect of the present invention, in the integrated processing means, the position of the template, the classification of the target of the template, the serial number of the template, and the image of the target existing at the position of the template are transmitted from the template processing means of each camera video processing system. The moving object extraction apparatus according to any one of claims 1 to 4, wherein the position of the moving object in the observation area is calculated using the estimated lower limit position information of the area in (1). When integrating information from each camera image processing system for achieving an effect, it has an effect that a large number and types of moving objects can be tracked.
[0012]
According to a sixth aspect of the present invention, in the integrated processing means, a mapping function of an installation position of each camera in an observation space, a pixel in an image, and a plane position of the observation space is known, and a template of each image processing system is provided. The center point A of the template extracted in the process and the estimated lower limit point B of the target area indicated by the template are plotted at a plane position in the observation space, and the points A and B are connected by a straight line. The moving object extraction device according to claim 1, wherein intersections of the straight lines obtained from a plurality of video processing systems are set as candidate positions of the target object. There is an effect that the position accuracy of the moving object can be improved by the integration processing means for performing the integration / identification processing of the moving object information.
[0013]
According to a seventh aspect of the present invention, the shortest distance processing in the observation space is performed by using information on the position candidate of the target object obtained from the integration processing means, the position of the target object obtained so far, and the target classification of the template. 7. The moving object extracting apparatus according to claim 6, wherein the association is performed in accordance with the time axis by the output processing means. The output processing means has the effect of reducing erroneous recognition of a moving object on the time axis for processing.
[0014]
The invention according to claim 8 of the present invention provides a template processing meansIsPerform two-dimensional filter processing from camera images,ExtractedEdge video dataToOn the basis ofGenerate a templateA moving object extraction apparatus according to any one of claims 1 to 7, wherein the template processing means according to the present invention effectively executes pattern matching processing by using edge information.
Has the effect of being able to
[0015]
The invention according to claim 9 of the present invention isShape information extraction meansCalculates the similarity of the target object from the number of pixels indicating the size of the region for each moving region extracted by the labeling processing means and the size of the region, and based on the similarity, a template indicating the target object GenerateLocation informationThe moving object extracting apparatus according to any one of claims 1 to 8, wherein the moving object is extracted, and has an effect that a large number and types of moving objects can be tracked in the present invention.
[0016]
The invention according to claim 10 of the present invention is:Shape information extraction meansThe degree of similarity extracted from is a moving object extraction apparatus according to claims 1 to 9, wherein the degree of similarity between a plurality of target objects such as a human being and a car is calculated. It has the effect of being able to track any type of moving object.
[0018]
Claims of the invention11In the invention described in (1), the template processing means includes a template number, a target object indicating what the target object was at the time of generating the template, a current position of the template, and a position update for indicating the attribute of the template when generating the template. Position, the number of updates indicating the number of updates, the disappearance flag indicating the disappearance when the template is to be erased, and the pixel value of TXS * TYS indicating the contents of the template (TXS, TYS is the size of the template. ).1The moving object extracting apparatus according to the present invention has an effect capable of giving flexibility to a large number of types, amounts, and processing in the present invention.
[0019]
Claims of the invention12In the invention described in (1), the template processing means generates the template in which the attribute of the template is set when the similarity extracted by the shape feature means is equal to or more than an appropriate threshold value.1 or 1The similarity from the shape feature means is used for generating template processing.
Thus, there is an effect that a template can be generated.
[0020]
Claims of the inventionThirteenAccording to the invention described in the above, the template processing means generates the template when the similarity extracted by the shape feature means isPredeterminedThe method according to claim 1, wherein when the value is equal to or more than the threshold value, the template is generated at a template setting position extracted by the shape feature means.1 or 11-12The moving object extraction device described above has an effect of outputting template generation position information other than the similarity from the shape feature means in the present invention, thereby enabling highly accurate template generation.
[0021]
Claims of the invention14According to the invention described in the above, in the template processing means, the similarity extracted by the shape feature means in the extracted moving region isPredeterminedIn the case where the value is equal to or larger than the threshold value, the periphery of the moving area is searched, and if the same template as the target object exists, the position of the template is updated and the pixel value of the position where the content of the template is moved is set to ExchangeRukoClaims characterized by the following:1 or 11-13The moving object extracting apparatus described above has an effect of preventing generation of a template overlapping one target object.
[0022]
According to a fifteenth aspect of the present invention, in the template processing means, each time a video signal is transferred from the input means, for a template whose position is not updated, template matching processing is performed centering on the current position of the template. And move the template to the position with the least error, and update the number of updates indicating the attribute of the template.Increase, The original positionpositionClaims characterized by performing setting1 or 11-14The moving object extraction device described above has an effect that the template matching method can be effectively used.
[0023]
Claims of the invention16According to the invention described in (1), the template processing means obtains a template motion vector after performing position update processing for all templates every time a video signal is transferred from the input means, and obtains the same motion vector. If there is more than one template, set a deletion flag for one template according to the position and the number of updates from among the multiple templates, andShow,If the template on which the extinction flag is set is continuously deleted, the template is deleted.1 or 11 to 15The moving object extracting apparatus described above has an effect of eliminating unnecessary templates and performing highly accurate template processing.
[0024]
An embodiment of the present invention will be described below with reference to FIGS.
(Embodiment 1)
FIG. 1 shows a block diagram of a moving object extracting apparatus according to Embodiment 1 of the present invention. In FIG. 1, reference numerals 7 to 9 denote camera image processing systems, 1 denotes an input means for inputting a camera image, 2 denotes a moving area extracting means for detecting a moving area from the video data obtained from the input means 1, and 3 denotes a moving image. Labeling processing means for detecting information on the moving area such as the size of the area and the number of pixels for each moving area from the moving areas obtained from the area extracting means 2, and the information on each moving area detected by the labeling processing means 3 The shape information extracting means 5 extracts information indicating how similar the moving area is to the target object, and the video data obtained from the input means 1 and the motion information obtained from the shape feature extracting means 4 are used. A template processing means for operating a template on the basis of area information; 6 an integrated processing means for integrating information from a plurality of camera video processing systems 7 to 9; Composed of output processing means for displaying Desupurei like to process the position and classification information of the moving object has.
[0025]
A large flow from the input means 1 to the template processing means 5, which are the camera image processing systems 7 to 9 of the present invention, is a process in which a target object discriminating method using shape features and a template matching method are combined. Template matching is a process that is stable against changes in the shooting environment and has a relatively light calculation load.However, when a template generation method and tracking of a moving object using the template are performed, Tracking becomes unstable with respect to size fluctuations. Therefore, in the present invention, the processing load is reduced by obtaining the similarity of the target object from the shape characteristics of the moving object, performing template generation / position update and template matching using the similarity, and tracking the moving object. , A highly accurate moving object template matching method.
[0026]
Hereinafter, each processing means will be described in detail.
The input means 1 uses, for example, a visible camera or an infrared camera as a camera. Although a plurality of cameras exist in the present invention, different types of cameras may be mixed even with only the same type of camera. Selection can be made according to the application.
[0027]
As an example of installation, as shown in FIG. 2, four observation areas are photographed by a plurality of cameras, provided that each area is photographed in an overlapping manner by two or more camera images. As one set, when these cameras are referred to as camera 1, camera 2, camera 3, and camera 4, cameras 1 and 2, and cameras 3 and 4 are arranged side by side or inward. There is also a method of increasing the number of cameras so that cameras 3 and 4 are installed opposite to each other so that any observation area can be photographed in an overlapping manner by two or more camera images. As the monitoring area in FIG. 2, an automobile road including an expressway is assumed.
[0028]
In the present invention, it is important for the moving area extraction processing means 2 to detect the entire target object, and it is preferable that the entire area including the foot is extracted as one area. When extracting a moving area, if the rate of separation of the area of the target object is small, a method for subtracting one frame of the video or a background video is calculated, and the difference from the current video using the background is calculated. The method used shall be used. Further, when the region of the target object is separated by the above-described method, an inter-frame cumulative difference is obtained by accumulating the difference result between one frame of the video over several frames, and then binarized with an optimal threshold. And a method of extracting a motion region.
[0029]
The labeling processing unit 3 divides the region of the target object (for example, a person or a car) from the binarized motion region detected by the moving region extraction unit 2 by a labeling process.What to doIt is. In general, the motion area detected by the motion area extraction unit 2 may include an area other than the target object, and examples thereof include a shadow, reflection, and noise added to the video. Therefore, the present means also has a role as an area shaping that removes these unnecessary parts and extracts a part close to the area of the actual target object as one area.
[0030]
The moving area detected by the labeling processing unit 3 of the shape information extraction processing unit 4 includes a case where the target object includes an overlap, separation, or a shadow. Therefore, when tracking a target object using a template described later in the present invention, a rule for generating a template for a new moving area is required. In the shape information extraction processing 4, “similarity” indicating the likelihood of determining whether or not the moving area is the target object to be tracked is calculated using the shape characteristics such as the size of the moving area, the Feret ratio, and the overall inclination. , A rule for making a determination is set. In the room, monitoring and tracking is performed for a person.
[0031]
In this case, there is a problem whether the moving area is a single person or another object (such as an overlapping or shadow of a plurality of persons). This can be determined from the fact that the actual size of the person is substantially determined. This determination rule will be described below.
[0032]
The size Size of the person on the image is related to the actual size by projective transformation via the distance from the camera.
[0033]
Therefore, this projective transformation is performed by the linear expression Size = axY.bottom  + B can be approximated. Where YbottomIs the coordinates of the lower end of the moving area on the image, and almost corresponds to the position of the foot of a person. The size having the length dimension includes the width w and height h of the circumscribed rectangle of the moving region, the square root r of the area (number of pixels) of the moving region, and the like. Each of w, h, and r can be well approximated by the above-described linear expression. Therefore, w, h, and r can be compressed into one quantity by principal component analysis. Then, Size is calculated as Size = αwxw + αhxh + αrxr. Where αw  , Αh, ΑrIs the projection coefficient onto the first principal component axis and αw 2+ Αh 2+ Αr 2= 1 is satisfied. Since Size can be well approximated by the above-described linear expression, the coefficients a and b of the above expression were obtained by an appropriate fitting method (least square method or the like), and “the standard size of the person at the foot” was determined by this expression.
[0034]
The discrimination rule was determined as follows using the standard size. The person similarity L in the moving area is calculated as L = 100 (1- | Size-Size0| / Size0), And a person is determined when L ≧ Th is satisfied. However, here, Size0Is the motion area YbottomStandard size (a xYbottom  + B), Th is an appropriate threshold.
[0035]
Note that the above shows the rules for a person.bottom  There is also a method in which a person or a car is determined based on the upper and lower relation of the approximate expression using a first-order approximate expression of the square root r of the area of the moving region.
[0036]
The template processing of the template processing means 5 uses the similarity calculated by the shape information extraction processing means 4 to generate a template and correct the position using the template creation candidate position when the similarity is high. If the similarity is low, template matching is performed to change the position of the template.
[0037]
FIG. 3 shows an example of template generation and position change. The specific template generation and position change are performed when the similarity of the detected moving area is equal to or more than an appropriate threshold (FIG. 3A). (FIG. 3 (2)), and a template is newly generated when there is no template in that range. If a template already exists in the search range, the closest template is moved to the update position designated by the shape information extraction processing means 4 (FIG. 3 (3)), and the pixel value of the template is moved to the updated position. All are replaced with pixel values.
[0038]
The template matching is performed on an edge image extracted from each frame image by Sobel filter processing. Therefore, the pixel value stored in the template is a value extracted from the edge image. In the template matching method, as a method of searching for a pattern similar to a template, there are a shortest distance method, an SSDA (Sequential Similarity Detection Algorithm), a correlation method, a statistical method, and the like. In the present invention, SSDA having high speed is used. However, in SSDA, when the pattern changes significantly, the template falls out of an area with few edges. To prevent this, the absolute value of the difference between the added value of the edge information in the template and the added value of the edge information of the reference pattern is added to the error when SSDA is executed.
[0039]
Further, in this means, there is the disappearance of the template. If there are a plurality of templates having the same movement vector calculated when the position is changed, the corresponding template is to be deleted.
[0040]
Next, FIG. 4 shows a block diagram of the inside of the template processing means 5. The template processing unit 5 includes an edge extraction processing unit 11, a template generation / position change processing unit 14, a template matching processing unit 12, an edge feature processing unit 13, and a template post-processing unit 15.
[0041]
The processing flow will be described below.
The edge extracting unit 11 performs an edge extracting process on the camera image 104 input from the input unit 1 for each frame. The information obtained by the edge extracting means is subjected to the above-described matching processing by the matching processing means 12 and edge feature processing by the edge feature processing means 13 in a range determined around the current position of the template. In the edge feature processing, as described above, matching between the edge pattern and the pattern of the template such as a process of performing a difference between the total value of the edge values in the template and the total value of the edge values of the region where the template of the edge image is matched is performed. Done. An error between the template matching processing and the edge feature processing is added by the template post-processing means 15, and the template is updated to a position where the added value is the smallest.
[0042]
The position and size information of the moving region and the similarity obtained from the shape information extraction processing means 4 are input to the template generation / position change processing means 14. The processing indicated by 3 is performed. If the similarity is larger than an appropriate threshold, the processing by this means is prioritized. Note that the template post-processing means 15 performs the above-described template disappearance processing.
[0043]
Next, FIG. 5 shows the structure of the template of the template processing means 5 in the present invention. When a template is created, it is created to have the following attributes and edge patterns where the template was created. That is, the template includes the serial number of the template, the current position of the template in the video, the position before the change, the target object (this is the identifier of the target object), the deduction flag, the still / moving flag, the number of updates, and the edge pattern. It is.
[0044]
Each attribute will be described below. Each template holds the coordinates on the current image and the previous coordinates. The reason why the previous coordinates are held is to obtain the movement amount vector. The movement amount vector is for finding a template that moves in the same direction, and templates that move continuously in the same direction are to be deleted. The number of updates is set to 0 at the time of generation, and increases by one each time it is updated. The number of updates is also a value indicating the strength of the template, and those having a low value are likely to be eliminated. The deduction point flag is turned on when there are two or more templates whose movement amount vectors move in the same direction, and the number of updates is reduced in the processing described later, or the deletion processing is performed. The static / dynamic flag is turned on for a template indicating an area in which the similarity is equal to or more than an appropriate threshold. The update rate of the template at the time of update is changed by the flag.
[0045]
A method for updating a template will be described. Basically, weighted addition of the pixel value of the previous template and the pixel value of the newly matched patternDo. hereIs the new pixel value of the coordinates in the template, is the previous pixel value of the template, and is the pixel value in the template of the pattern that the template matched.is there.
[0046]
I meanWhen the target object is close to a stationary state, the update amount is increased. When updating is performed, if the disappearance flag is OFF, the number of updates is incremented by one.
[0047]
Finally, the disappearance of the template will be described. When a plurality of templates having the same movement amount vector exist, the disappearance flag is set to ON. When the disappearance flag is ON, the number of updates is changed as follows when updating the template.
[0048]
When update count> T1 Update count = T2
When T1> update count> T2 Update count = T3
Update count<  For T3 disappeared
Here, T1, T2, T3Is T1>T2> T3 Have a relationship. Therefore, when the disappearance flag is set three times in succession, the template is erased. Note that the target object (target identifier) of the template is used when it is necessary to distinguish between a person and a vehicle, such as in an outdoor scene.
[0049]
The integration processing means 6 integrates the positions of the templates generated in the above-described processing by the camera video processing systems 7 to 9 by a multi-viewpoint method. A specific example is shown in FIG. In FIG. 6, the camera 1 and the camera 2 photograph substantially the same observation area from different angles. At this time, the video of the camera 1 and the video of the camera 2 track the moving object by template processing. This tracking is performed using a template, and the position of the template indicates a part of the target (the head in the case of a person). For this reason, even if it is known how the camera is installed in the real space, the position of the floor of the target object (foot position in the case of a person) cannot be accurately obtained. Therefore, as shown in the shape information extraction processing means 4, the height h of the person is the Y coordinate Y of the lower end of the moving area of the person.bottomCan be approximated by a linear expression, so from the center point of the template in each camera image, the estimated bottom position of the target is approximately estimated from this linear expression at the point where the center point extends perpendicular to the floor surface direction. can do.
[0050]
For example, in FIG. 6, the camera image 1 (FIG. 6A) corresponds to the center point A of the template, the estimated bottom point B of the object, and the camera image 2 (FIG. 6B) corresponds to A and B of the camera image 1. A ` and B `. If it is assumed that the camera installation position is known for each point extracted from each image, if the image distortion is not so large, it can be converted by a linear equation.
[0051]
The right side of FIG. 6 shows an image diagram (FIG. 6C) of the real space in which the target trace is performed. In this real space, a straight line connecting points A and B and points A ` and B ` is plotted in the real space. In this case, assuming that the template processing of each video is correctly performed, an intersection is set at the position of the corresponding target object in the real space, except when the target object is on a line connecting the two cameras. It can be proved geometrically to have. Therefore, this intersection can be set as the position of the target object in the real space.
[0052]
The output processing means 10 obtains the intersection by the integration processing means 6, and then plots the position in the real space while associating the position of the target object existing in the past with the shortest distance according to the time axis. Tracing of each target object can be performed. In the template matching process, there are cases where the position of the template does not always accurately indicate the position of the target object, but robustness is ensured by associating with the shortest distance. In addition, when an intersection is not determined or when a plurality of target objects exist, a fictitious intersection may occur, but the processing is performed by association using a past history.
[0053]
Next, specific examples of the present invention will be described below.
As an embodiment of FIG. 1, there is a system for monitoring a person indoors. At this time, it is often difficult to correctly extract and track a person with one camera. For example, there are cases where the floor area is large, the lower part of the target object is concealed by a desk or the like, or the target surface is reflected or shadowed on the floor surface. Therefore, two or more cameras are installed, and tracking of multiple moving objects is performed by integrating and identifying multi-viewpoint moving object information using an automatically generated template matching method. The system can be realized.
[0054]
Also, in an outdoor video surveillance system, there are many scenes where vehicles and people are mixed in a wide monitoring area. In this case as well, two or more cameras are installed, and tracking of a plurality of moving objects is performed by integrating and identifying multi-viewpoint moving object information using an automatically generated template matching method. , A simple system can be realized.
[0055]
Although the width is not so wide as shown in FIG. 2, in a surveillance system on a highway or a general road which extends horizontally, the surveillance areas are regularly overlapped while installing cameras as shown in FIG. A high-accuracy, simple system is realized by integrating and identifying multiple viewpoints of vehicle information using the automatically generated template matching method to track multiple vehicles using the captured images. it can.
[0056]
【The invention's effect】
As described above, in the present invention, the feature of moving image processing is achieved by integrating and identifying multi-view moving object information using an automatically generated template matching method for tracking a moving object in an automatic monitoring system or the like. By using simple processing, it is possible to track a moving object with high accuracy, and it is possible to easily obtain the position of the overlapping or stopping object by the automatically generated template matching method. It is possible to track a moving object with high accuracy.
[0057]
Further, it is possible to omit the difficulty and inefficiency of pattern matching found in stereo processing. In addition, it is possible to observe a wide area by dispersing the installation positions of the cameras.
[Brief description of the drawings]
FIG. 1 is a block diagram of a moving object extracting apparatus according to a first embodiment of the present invention.
FIG. 2 is a layout example of a plurality of cameras in the moving object extraction device according to the first embodiment of the present invention;
FIG. 3 is a diagram showing an example of searching for a template at the time of generation or position change in the template processing means of the moving object extraction device according to the first embodiment of the present invention;
FIG. 4 is a block diagram showing a template processing unit of the moving object extracting apparatus according to the first embodiment of the present invention;
FIG. 5 is a structural diagram of a template of the moving object extraction device according to the first embodiment of the present invention;
FIG. 6 is a diagram illustrating a processing example of an integrated processing unit of the moving object extraction device according to the first embodiment of the present invention.
[Explanation of symbols]
1 Input means
2 Moving area extraction means
3 Labeling processing means
4 Shape information extraction processing means
5 Template processing means
6 Integrated processing means
7, 8, 9 camera image processing system
10 Output processing means
11 Edge extraction processing means
12 Matching processing means
13 Edge feature processing means
14 Template generation / position change processing means
15 Template post-processing means

Claims (16)

  1. In an apparatus for inputting a plurality of camera images and tracking a moving object in an observation area which can be photographed by the camera images, an input unit for inputting a camera image, and a moving region is detected from image data obtained from the input unit. Moving region extracting means, labeling processing means for detecting information on the moving region of the size and the number of pixels for each moving region from the moving region obtained from the moving region extracting means, and the labeling processing means Shape information extracting means for extracting the similarity between the moving area and the target moving object by using the information of each moving area ; video data and the shape information obtained from the input means based on the similarity; the information of the moving region obtained from the extraction means, and updates the closest template if there is already a template in the search range, there is no template that range situ Generates a template, and integration processing unit for integrating the camera image processing system that includes a template processing means for outputting information of the template as the information of the moving object, the information from the plurality of camera images processing system, wherein A moving object extraction device comprising output processing means for processing and outputting the position and classification information of the moving object obtained from the integration processing means.
  2. 2. The moving object extracting apparatus according to claim 1, wherein each of the observation areas photographed by a plurality of cameras is photographed in an overlapping manner by two or more camera images.
  3. When a plurality of cameras are installed as a set of four cameras, and these cameras are camera 1, camera 2, camera 3, and camera 4, camera 1 and camera 2 and camera 3 and camera 4 are arranged in parallel or inside. The cameras 1 and 2 and the cameras 3 and 4 are installed facing each other in the direction toward, and the number of cameras is increased so that any observation area can be overlapped with two or more camera images. The moving object extraction device according to claim 1.
  4. 4. The moving object extraction device according to claim 1, wherein the input unit is a visible camera, an infrared camera, or a combination thereof.
  5. The integrated processing means uses the template position, template object classification, template serial number, and the estimated lower limit position information of the area in the target image existing at the template position from the template processing means of each camera image processing system to obtain the observation area. The moving object extraction device according to claim 1, wherein the position of the moving object is calculated.
  6. In the integrated processing means, a mapping function of the installation position of each camera in the observation space, a pixel in the video, and a plane position of the observation space is known, and the center point A of the template extracted in the template processing of each video processing system and its The estimated lower limit point B of the target area indicated by the template is plotted at a plane position in the observation space, and the points A and B are connected by a straight line. 6. The moving object extracting apparatus according to claim 1, wherein intersections of the straight lines obtained from a plurality of video processing systems are set as candidate positions of the target object.
  7. Outputs that the correspondence is performed according to the time axis in the shortest distance processing in the observation space using the information on the position candidate of the target object obtained from the integrated processing means and the position of the target object obtained so far and the target classification of the template. 7. The moving object extraction device according to claim 6, wherein the processing is performed by processing means.
  8. Template processing means performs two-dimensional filtering the camera image, the extracted moving object extraction device according to any one of claims 1 to 7, characterized in that to generate a template based on the edge image data.
  9. The shape information extracting means calculates the similarity of the target object from the number of pixels indicating the size of the area for each moving area extracted by the labeling processing means and the size of the area, and based on the similarity, moving the object extracting apparatus according to any one of claims 1 to 8, characterized in that extracts the position information that generates a template that indicates the object.
  10. The moving object extraction device according to claim 1, wherein the similarity extracted from the shape information extracting unit calculates a similarity between a plurality of target objects such as a human and a car.
  11. The template processing means includes a template number, a target object indicating what the target object was at the time of generating the template, a current position of the template, and an original position when the position is updated to indicate the attribute of the template when generating the template. , The number of updates indicating the number of updates, the disappearance flag indicating the disappearance when the template is the object of disappearance, and a pixel value of TXS * TYS indicating the contents of the template (TXS, TYS indicates the size of the template). 2. The moving object extracting apparatus according to claim 1, wherein
  12. The template processing unit, generating the template shape information similarity extracted from the extraction means according to claim 1 or 11 further characterized in that to produce a template with the set of attributes of the templates if more than a predetermined threshold value Moving object extraction device.
  13. In the template processing means, when the similarity extracted by the shape information extracting means is equal to or more than a predetermined threshold, the template generating means generates a template at an appropriate position to be generated by the template at the template setting position extracted by the shape information means. moving the object extracting apparatus according to claim 1, 1 1 or 12, characterized in that to.
  14. When the similarity extracted by the shape information extracting means in the extracted moving area is equal to or more than a predetermined threshold, the template processing means searches around the moving area and finds the same template as the target object. when updates the position of the template, moving object extraction device according to any one of claims 1 or 11 to 13, wherein the Turkey interchanging the pixel value of the position moved the contents of the template.
  15. In the template processing means, each time a video signal is transferred from the input means, for a template for which the position of the template is not updated, a template matching process is performed centering on the current position of the template, and the template is positioned at a position having the least error. 15. The moving object extracting apparatus according to claim 1 , wherein the moving object extracting apparatus moves , increases the number of updates indicating the attribute of the template, and sets the position before the moving to the original position. .
  16. In the template processing means, each time a video signal is transferred from the input means, after performing position update processing for all templates, a motion vector of the template is obtained, and a plurality of templates having the same motion vector are obtained. If it exists, in accordance with the position and the number of updates from among the plurality of templates, make a annihilation flag on one of the templates, indicates that became the erasing target, annihilation flag of this is standing template 16. The moving object extracting apparatus according to claim 1, wherein the template is erased when is continuously deleted .
JP05132397A 1997-03-06 1997-03-06 Moving object extraction device Expired - Fee Related JP3577875B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP05132397A JP3577875B2 (en) 1997-03-06 1997-03-06 Moving object extraction device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP05132397A JP3577875B2 (en) 1997-03-06 1997-03-06 Moving object extraction device

Publications (2)

Publication Number Publication Date
JPH10255057A JPH10255057A (en) 1998-09-25
JP3577875B2 true JP3577875B2 (en) 2004-10-20

Family

ID=12883719

Family Applications (1)

Application Number Title Priority Date Filing Date
JP05132397A Expired - Fee Related JP3577875B2 (en) 1997-03-06 1997-03-06 Moving object extraction device

Country Status (1)

Country Link
JP (1) JP3577875B2 (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003044859A (en) * 2001-07-30 2003-02-14 Matsushita Electric Ind Co Ltd Device for tracing movement and method for tracing person
KR100790860B1 (en) * 2004-12-14 2008-01-03 삼성전자주식회사 Human tracking apparatus and method thereof, and recording media and removing electronic system therefor
JP2006209698A (en) * 2005-01-31 2006-08-10 Olympus Corp Target tracking device, microscope system and target tracking program
WO2007099762A1 (en) * 2006-03-01 2007-09-07 Nikon Corporation Object-seeking computer program product, object-seeking device, and camera
KR101030430B1 (en) * 2007-09-12 2011-04-20 주식회사 코아로직 Apparatus and method for processing image and computer readable medium stored thereon computer executable instruction for performing the method
JP5147760B2 (en) * 2009-03-02 2013-02-20 セコム株式会社 Image monitoring device
JP5147761B2 (en) * 2009-03-02 2013-02-20 セコム株式会社 Image monitoring device
US20120128209A1 (en) * 2009-06-30 2012-05-24 Shigetoshi Sakimura Image analysis device and image analysis program
CN102414719B (en) * 2009-07-22 2014-10-15 欧姆龙株式会社 Surveillance camera terminal
JP5018932B2 (en) 2010-06-23 2012-09-05 株式会社ニコン Imaging device
WO2013172491A1 (en) * 2012-05-16 2013-11-21 전자부품연구원 Method and device for recognizing object by using depth information
JP2014016839A (en) * 2012-07-09 2014-01-30 Nippon Telegr & Teleph Corp <Ntt> Mobile object tracking device
JP6276519B2 (en) * 2013-05-22 2018-02-07 株式会社 日立産業制御ソリューションズ Person counting device and human flow line analyzing device
JP6481960B2 (en) * 2015-03-23 2019-03-13 日本電気株式会社 Information processing system, information processing method, and program for detecting position of foot

Also Published As

Publication number Publication date
JPH10255057A (en) 1998-09-25

Similar Documents

Publication Publication Date Title
US10043315B2 (en) Method and apparatus for representing a virtual object in a real environment
US10366538B2 (en) Method and device for illustrating a virtual object in a real environment
US20180357786A1 (en) Method of Providing a Descriptor for at Least One Feature of an Image and Method of Matching Features
US20180341823A1 (en) Processing method for distinguishing a three dimensional object from a two dimensional object using a vehicular system
Teoh et al. Symmetry-based monocular vehicle detection system
US9417310B2 (en) Airport target tracking system
Jazayeri et al. Vehicle detection and tracking in car video based on motion model
JP6295645B2 (en) Object detection method and object detection apparatus
Tardif et al. Monocular visual odometry in urban environments using an omnidirectional camera
Senior et al. Appearance models for occlusion handling
Beymer et al. Real-time tracking of multiple people using continuous detection
Smith et al. ASSET-2: Real-time motion segmentation and shape tracking
US5436672A (en) Video processing system for modifying a zone in successive images
US6542621B1 (en) Method of dealing with occlusion when tracking multiple objects and people in video sequences
CA2326816C (en) Face recognition from video images
Sidla et al. Pedestrian detection and tracking for counting applications in crowded situations
US10417503B2 (en) Image processing apparatus and image processing method
US7321386B2 (en) Robust stereo-driven video-based surveillance
Toulminet et al. Vehicle detection by means of stereo vision-based obstacles features extraction and monocular pattern analysis
EP2602761A1 (en) Object detection device, object detection method, and program
EP0567059B1 (en) Object recognition system using image processing
Yoneyama et al. Moving cast shadow elimination for robust vehicle extraction based on 2D joint vehicle/shadow models
US6701005B1 (en) Method and apparatus for three-dimensional object segmentation
US8170278B2 (en) System and method for detecting and tracking an object of interest in spatio-temporal space
US6911995B2 (en) Computer vision depth segmentation using virtual surface

Legal Events

Date Code Title Description
A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20040206

A131 Notification of reasons for refusal

Free format text: JAPANESE INTERMEDIATE CODE: A131

Effective date: 20040217

A521 Written amendment

Free format text: JAPANESE INTERMEDIATE CODE: A523

Effective date: 20040331

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20040622

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20040705

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20070723

Year of fee payment: 3

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20080723

Year of fee payment: 4

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20090723

Year of fee payment: 5

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20090723

Year of fee payment: 5

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20100723

Year of fee payment: 6

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20110723

Year of fee payment: 7

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20110723

Year of fee payment: 7

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120723

Year of fee payment: 8

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120723

Year of fee payment: 8

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20130723

Year of fee payment: 9

LAPS Cancellation because of no payment of annual fees