JP2007227992A - Object detecting system, object detecting method, monitoring method an monitoring system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an object detecting system and a method thereof capable of improving the efficiency and accuracy of target selection, and increasing the interference of environmental variables. <P>SOLUTION: The object detecting system is provided, which includes a processor for differentiating a target area into at least a first area and a second area, a controller for designating a first weight for the first area and a second weight for the second area, and a first video capture member for first capturing video images of a target object appearing in the first area, before capturing video images of a target object appearing in the second area. The system can allow the first weight to be higher than the second weight. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an object detection system and method that uses multiple area and weight designation schemes in an active surveillance system to increase the efficiency and accuracy of target selection and reduce environmental variable interference.

  Most surveillance monitors and their systems currently in use are limited in that the resolution and the proportion of the target object on the screen are too small. That is, since the resolution is too low, the screen is not clear, and even the screen obtained through the digitization process cannot effectively improve this problem. For this reason, an affiliated company has developed an active video tracking system, that is, using a video capture member to obtain panoramic screen data, and further comparing the video with a processor to obtain a target object, and another video. By driving the capture member to obtain partial screen data, it is possible to effectively acquire screen data with better quality of the target object. This effectively improves the quality of the screen, but there are not a few problems in practice. For example, on the panoramic view screen, it is difficult to acquire a target object when all the areas have different importance such as people and cars. In addition, interference with the outside environment (trees, flags, shadows, etc. that move in the wind) may complicate the environment, and multiple target objects may appear at the same time. Cannot be secured.

  Furthermore, since the conventional monitoring monitor stores the video data of the entire view, the video data becomes large. Therefore, in order to lengthen the monitoring recording time, a long recording time is ensured with a monochrome screen and a low resolution. For this reason, how to find a target object, determine its importance, and improve the recording quality and efficiency is an important problem that needs to be solved.

  In view of the above-mentioned drawbacks of the conventional monitoring monitor and its system, the inventor of the present invention, based on the experience accumulated over many years in the related industry, to research and overcome the imperfection. In the active monitoring system of the present invention, a method of assigning multiple areas and weights is used to improve the efficiency and accuracy of target selection and to prevent the interference of environmental variables. Can be reduced.

  The main object of the present invention is to use a method of specifying multiple areas and weights in an active monitoring system to increase the efficiency and accuracy of target selection and reduce the interference of environmental variables It is to provide a detection system and method.

  Another object of the present invention is to provide an object detection system and method that can be enlarged several times when capturing video data of a specific target object, and can display a high-quality and clear video screen. There is to do.

  Further, another object of the present invention is to process the background and the target object in each area, reduce the resolution of the background, and increase the resolution of the target object. It is an object of the present invention to provide an object detection system and method capable of reducing the amount of occupation and storing useful video data in a memory.

  To achieve the above objective, the object detection system and method of the present invention is used to detect at least one target object appearing in a target area by an active monitoring monitor, which system and method mainly include: A processor used to divide the target area into at least a first area and a second area; a controller used to designate the first weight as the first area and the second weight as the second area; and the second area Before capturing the video of the target object that appears in the first area, first includes a first video capture member that captures the video of the target object that appears in the first area, the first weight is higher than the second weight .

  This makes it possible to increase the efficiency and accuracy of target selection and reduce the interference of environmental variables by using a method of specifying multiple areas and weights in an active monitoring system.

  Hereinafter, the present invention will be described in detail based on specific embodiments and drawings in order to facilitate understanding of the objects, features, and effects of the present invention.

  The object detection system and method of the present invention is used to detect at least one target object appearing in a target area, and mainly includes the following members.

  A processor used to divide the target area into at least a first area and a second area. The processor can change the size and position of the first zone and the second zone in a predetermined time period.

  A controller used to designate a first weight as a first zone and a second weight as a second zone. Here, the first weight is set higher than the second weight. The controller can change the first weight and the second weight at a constant predetermined time period, and the values of the first weight and the second weight increase at least in a predetermined direction.

  A video capture member that first captures an image of a target object that appears in the first area before capturing an image of the target object that appears in the second area. A second video capture member that captures a panoramic video of the target area may be further included. Note that the video viewing angle of the first video capture member is smaller than that of the second video capture member.

  FIG. 1 shows the best embodiment of the present invention. As shown in FIG. 1, the convenience store is divided into a first area 1 of an entrance / exit area, a second area 2 of a customer area, and a third area 3 of a counter area, which are used for capturing an image of a target object ( The target object is selected as a moving object). Then, different weights are designated for each area, such as first area 1> second area 2> third area 3, and the image capturing member captures the area with the highest weight first. Therefore, when there are target objects in all the areas, the target objects in the important areas are preferentially captured.

  For example, the first image capturing member preferentially captures an image of a person passing through the entrance / exit area of the first area 1. If no one passes through the first zone 1 for a predetermined time, the processor reduces the weight of the first zone 1 and the first video capture member appears in the customer area of the second zone 2. Capture the video of the target object with priority. Furthermore, the video of the target object in the third zone 3 is captured only when there are no people in the first zone 1 and the second zone 2.

  2 and 3 show the best embodiment of the present invention. As shown in these figures, when the present invention is used for general target object tracking outdoors, the shadow of a tree that blows and shakes in the wind or a tree whose length changes due to a change in the position of the sun is the target of the image capture member. Very often affects object selection. Therefore, in order to avoid the occurrence of the above situation, the interference of the external environment is suppressed by reducing the weight of the shadow 4 of the area 4 that changes with the tree 4 of the fixed area and the time, or zero, The efficiency of target selection can be improved.

  4 and 5 show the best embodiment of the present invention. In these figures, a first video capture member and a second video capture member are used. The second video capture member mainly captures a panoramic video, has a large viewing angle, and a large capture range. On the other hand, the first video capture member mainly tracks and captures the target object, and its viewing angle is small and the range to be captured is small, so that the resolution of the target object can be improved. When capturing the image of the entire area, the second image capture member divides the entire view into several areas and determines different weight levels depending on the perspective and direction of the field of view. Therefore, the size of the distant target object is too small to be overlooked, and the probability of being selected even in the distant target object is the same. The size of the image of the same target object varies depending on the distance, but the probability that the target object is selected at each position becomes equal by adjusting the weight level.

  As shown in FIG. 6, when the present invention is applied to road monitoring, the capture screen is divided into a first area 6, a second area 7, a third area 8, and a fourth area 9, each of which is a human activity area. Corresponding to two-way traffic roads and forest areas. And by setting the weights as second zone 7 and third zone 8 (two-way roadway)> first zone 6 (human activity zone)> fourth zone 9 (forest zone) An enlarged image of the target object to be monitored can be acquired.

  In order to effectively improve the efficiency of storing video data in the memory, the present invention can be applied to a monitoring system, and the system includes the following members.

  The video capture device is used to acquire first panoramic video data having a first resolution from a predetermined monitoring area.

  The first data storage member can receive the first panoramic video data captured by the video capture device, store the first panoramic video data, and transmit the first panoramic video data to a data identification member described later.

  The data identification member is connected to the first data storage member, receives the first panoramic video data transmitted from the first data storage member, and identifies video data of at least one target object from the first panoramic video data. In addition to identifying video data, the first panoramic video data can be divided into target object video data and background video data. The data identification member includes an object detection system used to detect at least one target object appearing in the panoramic video data, and the object detection system includes the following members.

  The processor is used to divide the panoramic video data into at least a first area and a second area.

  The controller is used to designate the first weight as the first zone and the second weight as the second zone, where the first weight is higher than the second weight.

  The image capturing member first captures an image of the target object appearing in the first area before capturing an image of the target object appearing in the second area.

  The data conversion member receives background video data from the data identification member and is used to convert the background video data from the first resolution to the second resolution, and the second resolution is lower than the first resolution.

  The video synthesizer is used to receive background video data having a second resolution and target object video data having a first resolution, and to form second panoramic video data. The second panoramic video data includes background video data having a second resolution and target object video data having a first resolution.

  7 and 8 show the best embodiment of the present invention. As shown in these figures, when the present invention is applied to road vehicle monitoring, the above-described monitoring system design is used, and the following procedure is included.

  Data of the first panoramic video 10 having the first resolution of the predetermined monitoring area is acquired by the video capture device.

  The data of the first panoramic image 10 is stored in the first data storage member, and the data of the first panoramic image 10 is transmitted from the first data storage member to the data identification member.

  The data identification member identifies video data of at least one target object 11 (car) from the data of the first panoramic video 10, and the data of the first panoramic video 10 is used as video data and background video data of the target object 11. Divide into Further, the data identification member includes an object detection system used for detecting at least one target object 11 appearing in the first panoramic video 10 data. The object detection system includes the following configuration.

  Data of the first panoramic image 10 is divided into at least a first area 101 and a second area 102. The first weight is designated as the first area 101 and the second weight is designated as the second area 102. Here, since the first weight is higher than the second weight, the image capturing member captures the target object appearing in the first area 101 before capturing the image of the target object appearing in the second area 102. 11 images are captured. The data conversion member receives the background video data from the data identification member, and converts the background video data from the first resolution to the second resolution. At this time, the second resolution is set lower than the first resolution. Subsequently, the video synthesizer receives the background video data having the second resolution and the target object video data having the first resolution to form data of the second panoramic video 12. Therefore, the data of the second panoramic video 12 includes background video data having the second resolution and video data of the target object 11 having the first resolution.

  A panoramic image 10 (including the target object) is acquired by the high-resolution video capture device, and further, through the video processing, the high resolution of the target object 11 in the panoramic image 10 is maintained, and all the video other than the target object 11 is displayed. By considering it as a background, the resolution reduction processing is performed. Therefore, the burden of subsequent data transmission and storage can be reduced, the amount of video data in the storage capacity of the memory can be reduced, and useful video data can be stored in the memory.

  The present invention can be applied to still another monitoring system, and this monitoring system includes the following configuration.

  The video capture device is used to acquire panoramic video data of a predetermined surveillance area.

  The first data storage member can receive the panoramic video data captured by the video capture device, store the panoramic video data, and then transmit the data to a data identification member to be described later.

  The data identification member is connected to the first data storage member, receives the first panoramic video data transmitted from the first data storage member, and identifies video data of at least one target object from the first panoramic video data. Further, the data identification member can divide the panoramic video data into target object video data and background video data. The data identification member includes an object detection system used to detect at least one target object appearing in the panoramic video data, and the object detection system includes the following configuration.

  The processor is used to divide the panoramic video data into at least a first area and a second area.

  The controller is used to assign a first weight to the first zone and a second weight to the second zone. Note that the first weight is higher than the second weight.

  The image capturing member first captures an image of the target object appearing in the first area before capturing an image of the target object appearing in the second area.

  The second data storage member receives the target object video data, stores the target object video data in the second data storage member, and then deletes the background video data in the first data storage member.

  9 and 10 show the best embodiment of the present invention. As shown in these drawings, data of a panoramic video 13 in a predetermined surveillance area is acquired using a video capture device.

  First, the data of the panoramic image 13 is stored in the first data storage member, and then transmitted from the first data storage member to the data identification member.

  Next, video data of at least one target object is identified from the data of the panoramic image 13 by the data identification member connected to the video capture member. Here, the video data 141 and 142 are identified.

  Then, the data of the panoramic video 13 is divided into video data of the target objects 141 and 142 and background video data. Here, the data identification member includes an object detection system used to detect at least one target object 141 appearing in the panoramic video 13 data. The detection system for the target object 141 includes the following configuration.

  The panoramic video 13 data is divided into at least a first area 131 and a second area 132. The first weight is designated as the first area 131, the second weight is designated as the second area 132, and the first weight is higher than the second weight.

  Next, before capturing an image of the target object 142 appearing in the second area 132, first, an image of the target object 141 appearing in the first area 131 is captured by the image capture member.

  Then, after transmitting the video data of the target object 141 to the second data storage member, the background video data in the first data storage member is deleted.

  In the next cycle, the image of the target object 142 is captured, only the images of the target objects 141 and 142 are stored, and all the backgrounds are deleted. The high-resolution video capture device captures the panoramic image 13 (including the target object), and further holds the video of the target objects 141 and 142 in the panoramic video 13 through video processing, and the video other than the target objects 141 and 142 All are considered background. Since all the background is deleted, the burden of subsequent data transmission and storage is reduced, the occupation of the video data in the storage capacity of the memory is reduced, and useful video data can be stored in the memory.

  The application of the present invention to still another monitoring system and its method are shown in FIG.

  The first video capture device 15 has a first video viewing angle, and can acquire data of a panoramic video 151 in a predetermined surveillance area.

  The second video capture device 16 is movable and has a second resolution. Then, it is possible to track at least one target object 17 appearing in a predetermined monitoring area and acquire the video. Note that the second video capture device 16 has a second video viewing angle smaller than the first video viewing angle.

  The object detection system is used for target object tracking control of the second video capture device 16 and includes the following members.

  The processor receives the panoramic video data from the first video capture device and divides the panoramic video 151 data into at least a first zone 152 and a second zone 153.

  The controller assigns the first weight to the first zone 152 and the second weight to the second zone 153. Here, the first weight is assumed to be higher than the second weight.

  The target object positioning member first determines the position of the image of the target object 17 appearing in the first area 152 before determining the position of the image of the target object appearing in the second area 153.

  The second video capture device 16 tracks the target object 17 according to the positioning of the target object 17 provided by the object detection system.

  The first video capture device 15 acquires the panoramic image 151, finds the position of the target object 17 via the object detection system, and further drives the second video capture device 16 to track the target object 17 for high resolution. Get video data.

  As described above, the design of the present invention is more innovative than the conventional video data storage system, has practicality, high usage characteristics, value, inventive step and novelty, and is completely patent-pending. In order to meet the statutory requirements of, we apply here for a patent under the law.

  Although the best embodiments of the present invention have been described above, these do not limit the scope of the present invention. All changes and modifications that do not depart from the spirit and scope of the present invention by those who are familiar with the related art are regarded as changes and modifications that obtain equivalent effects based on the present invention, and are included in the patent scope of the present invention. The scope of the present invention is in accordance with the scope of the claims.

It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention. It is a figure which shows the Example of this invention.

Explanation of symbols

1 first zone
2 Second area 3 Third area
4 trees 5 shadows
6 1st area 7 2nd area
8 Third area 9 Fourth area
10 First panoramic video 101 First area
102 Second area 11 Target object
12 Second panoramic video 13 panoramic video
131 1st area 132 2nd area
141 Target object 1
142 Target object 2
15 First video capture device 151 panoramic video
152 1st area 153 2nd area
16 Second video capture device 17 Target object

Claims (26)

An object detection system used for detecting at least one target object appearing in a target area,
The system includes a processor that divides the target area into at least a first area and a second area;
A controller designating a first weight as the first zone and a second weight as the second zone;
A first video capture member that first captures an image of the target object that appears in the first area before capturing an image of the target object that appears in the second area;
The object detection system according to claim 1, wherein the first weight is higher than the second weight.   The object detection system according to claim 1, further comprising a second video capture member that captures a full-view video of the target area.   The object detection system according to claim 2, wherein a video viewing angle of the first video capture member is smaller than a video viewing angle of the second video capture member.   2. The object detection system according to claim 1, wherein when the second weight is set to 0, the first image capturing member does not capture an image of the target object appearing in the second area. .   The object detection system according to claim 1, wherein the processor changes a size and a position of the first area and the second area in a predetermined time period.   The object detection system according to claim 1, wherein the controller changes the first weight and the second weight at a predetermined time period.   The first video capture member captures the target object appearing in the first zone or the second zone, and after a predetermined time, the processor lowers the first weight or the second weight The object detection system according to claim 1.   The object detection system according to claim 1, wherein the values of the first weight and the second weight increase at least in a predetermined direction. An object detection method used for detecting at least one target object appearing in a target area, wherein the method divides the target area into at least a first area and a second area,
A first weight is designated as the first zone, a second weight is designated as the second zone, and the first weight is higher than the second weight;
An object detection method comprising: first capturing an image of the target object in the first area with a first image capture member before capturing an image of the target object in the second area. .   The object detection method according to claim 9, further comprising: a procedure in which a second video capture member captures a full-view video of the target area.   The object detection method according to claim 10, wherein a video viewing angle of the first video capture member is smaller than the second video capture member.   The object detection according to claim 9, wherein when the second weight is set to 0, the first image capturing member does not capture an image of the target object appearing in the second area. Method.   The object detection method according to claim 9, further comprising changing a size and a position of the first area and the second area at a predetermined time period.   The object detection method according to claim 9, further comprising a step of changing the first weight and the second weight at a predetermined time period.   The first image capturing member captures the target object appearing in the first area or the second area, and lowers the first weight or the second weight after a predetermined time. The object detection method according to claim 9, further comprising:   The object detection method according to claim 9, wherein the values of the first weight and the second weight increase at least in a predetermined direction.   A video capture device, a first data storage member, and a data identification member, wherein the video capture device acquires first panoramic video data having a first resolution from a predetermined monitoring area, and the first data storage member Receives the first panoramic video data captured by the video capture device, and the first panoramic video data is stored in the first data storage member and then transmitted to the data identification member, and the data An identification member is connected to the first data storage member, receives the first panoramic video data transmitted from the first data storage member, and receives video data of at least one target object from the first panoramic video data And the first panoramic video data can be divided into target object video data and background video data. An object detection system for detecting at least one target object appearing in video data, the object detection system further including a processor, a controller, a video capture member, a data conversion member, and a video synthesizer. The processor divides the panoramic video data into at least a first area and a second area, and the controller designates a first weight as the first area and a second weight as the second area. The first weight is higher than the second weight, and the image capturing member first appears in the first area before capturing an image of the target object appearing in the second area. A video of a target object is captured, and the data conversion member receives the background video data from the data identification member and receives the background video. Data is converted from the first resolution to the second resolution, the second resolution is lower than the first resolution, and the video synthesizer converts the background video data having the second resolution and the first resolution to the second resolution. A monitoring system, comprising: receiving the target object video data, and forming second panoramic video data including the background video data having the second resolution and the target object video data having the first resolution.   First panoramic video data having a first resolution is obtained from a predetermined monitoring area by a video capture device, the first panoramic video data is stored in a first data storage member, and the first panoramic video data is stored in the first panoramic video data. A data storage member to transmit to the data identification member, the data identification member identifies video data of at least one target object from the first panoramic video data, and the first panoramic video data is the target object video data and The image is classified into background video data, and the data identification member includes an object detection system that identifies at least one target object appearing in the first panoramic video data, and the object detection system further includes the panoramic video data. Dividing into at least a first zone and a second zone, designating a first weight as the first zone and a second weight as the second zone The first weight is higher than the second weight, and before capturing the image of the target object appearing in the second area, first appearing in the first area with a video capture member A video of a target object is captured, the background video data is received from the data identification member by a data conversion member, the background video data is converted from the first resolution to the second resolution, and the second resolution is the first resolution The background video data having the second resolution and the target object video data having the first resolution are received by a video synthesizer, and the background video data and the first resolution having the second resolution are received. A monitoring method comprising a step of forming second panoramic video data including the target object video data.   A video capture device, a first data storage member, and a data identification member; the video capture device acquires panoramic video data of a predetermined surveillance area; and the first data storage member Receiving the panoramic video data to be captured, storing the panoramic video data, and transmitting the data to the data identification member; the data identification member being connected to the first data storage member; Receiving the first panoramic video data to be transmitted, identifying video data of at least one target object from the first panoramic video data, and dividing the panoramic video data into target object video data and background video data; The data identification member is used for detecting at least one target object appearing in the panoramic video data. An object detection system, wherein the object detection system further includes a processor, a controller, a video capture member, and a second data storage member, wherein the processor converts the panoramic video data into at least the first area and the second area. The controller assigns a first weight to the first area and a second weight to the second area, the first weight being higher than the second weight; The capture member captures an image of the target object appearing in the first area before capturing the image of the target object appearing in the second area, and the second data storage member After receiving the target object video data and storing the target object video data in the second data storage member, in the first data storage member Monitoring system the background image data is characterized in that it is removed that.   A panoramic video data is acquired from a predetermined monitoring area by a video capture device, the panoramic video data is stored in a first data storage member, and the panoramic video data is transmitted from the first data storage member to a data identification member. The video identification member connected to the video capture member identifies video data of at least one target object from the panoramic video data, divides the panoramic video data into target object video data and background video data, and the data identification The member includes an object detection system that detects at least one target object appearing in the panoramic video data, and the object detection system further divides the panoramic video data into at least a first zone and a second zone, One weight is designated as the first zone and a second weight is designated as the second zone, and the first weight is designated as the first zone. Before capturing the video of the target object appearing in the second area, and first capturing the video of the target object appearing in the first area with a video capture member. A monitoring method comprising: capturing, transmitting the target object video data in a second data storage member, and deleting the background video data in the first data storage member after transmission.   A first video capture device, a second video capture device, and an object detection system, wherein the first video capture device has a first resolution and acquires panoramic video data from a predetermined surveillance area. The second video capture device has a second resolution and is capable of tracking and acquiring at least one target object appearing in the predetermined surveillance area. The object detection system controls tracking of the second video capture device, the object detection system including a processor, a controller, and a target object positioning member, wherein the processor Receiving the panoramic video data from the video capture device, dividing the panoramic video data into at least a first zone and a second zone; The controller designates a first weight as the first area and a second weight as the second area, the first weight is higher than the second weight, and the target object positioning member is Before determining the position of the image of the target object appearing in the second area, first, the position of the image of the target object appearing in the first area is determined, and the second image capture device A monitoring system that tracks the target object based on positioning of the target object provided by a detection system.   The monitoring system according to claim 21, wherein the second video capture device is movable.   The monitoring according to claim 21, further comprising a movable mirror, wherein when the second video capture device cannot be moved, the video of the target object is reflected to the second video capture device. system.   A panoramic video data is acquired from a predetermined surveillance area by a first video capture device having a first video viewing angle, and a motion of at least one target object appearing in the predetermined surveillance area is detected, and the detection A method receives the panoramic video data from the first video capture device, divides the panoramic video data into at least a first zone and a second zone, designates a first weight as the first zone, and a second A weight is designated as the second area, wherein the first weight is higher than the second weight, and before the position of the image of the target object appearing in the second area is determined, Based on the positioning of the target object provided by the object detection system by a second video capture device, determining a position of the video of the target object appearing in one area Serial comprises steps of tracking a target object, of which the monitoring method the second image capturing device is characterized in that it comprises the steps of, with the first scene image viewing angle is smaller than the second scene image viewing angle.   The monitoring method according to claim 24, wherein the second video capture device is movable.   25. When the second video capture device is immovable, the target object is tracked by a movable mirror, and an image of the target object is reflected to the second video capture device. Monitoring method described in 1.

Images