JPH04261290A - Moving object tracing device - Google Patents

Abstract

PURPOSE:To minimize the effect such as image pickup environment by minimizing a time delay as the result of change area extraction and extracting a moving object at all times even when the moving object enters a monitor area and comes to a standstill. CONSTITUTION:A change area extraction section 3 uses an inputted picture continuous in time series to extract a change area of a moving object. A rectangular information acquisition section 4 acquires the rectangular information including the extracted change area and magnifies the acquired rectangular information into rectangular information with a size of a prescribed magnification. A projection processing section 5 applies picture processing to an area corresponding to the inputted picture continuous in time series to trace the moving object based on the magnified rectangular information.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving object tracking device for tracking a moving object in an image monitoring device for detecting illegal intruders using images captured by a television camera, for example.

0002

[Background Art] In recent years, image monitoring devices that detect illegal intruders, etc. by using images captured by television cameras (hereinafter referred to as television cameras) have become widespread, for example in security systems. . This is because recent remarkable developments in semiconductor and computer system technology and the widespread use of digital image processing technology have made it possible to create low-cost devices.

In such an image monitoring device, when detecting and tracking a moving object, a changing region of the moving object is extracted by differential processing between images that are continuously inputted in time series, and a changing region of this changing region is extracted. Information such as area is used to determine whether the object is a moving object (invading object). The moving object is tracked by repeating the above process.

0004

[Problems to be Solved by the Invention] However, in the means for detecting and tracking moving objects in conventional image monitoring devices, as described above, by always performing subtraction processing between a plurality of consecutive images, I'm trying to extract a region of change. For this reason, the final change area extraction result is delayed in time. If an attempt is made to shorten this delay time, the sampling time of the input image will be shortened, making it impossible to obtain sufficient change information. Furthermore, since processing is always performed on a plurality of input images, if a moving object once extracted becomes stationary, this moving object cannot be extracted. Furthermore, it becomes extremely difficult to extract a changing area due to changes in the photographing environment of the television camera. For example, there has been a problem in that it becomes easier to extract noise areas depending on wind conditions and sunlight conditions.

Therefore, the present invention can minimize the time delay in the result of extracting a changed region, can always extract a moving object even if it comes to a standstill after entering the monitoring region, and can An object of the present invention is to provide a moving object tracking device that can minimize the influence of the environment.

[0006]

[Means for Solving the Problems] A moving object tracking device of the present invention includes an image input means for inputting images continuously in chronological order, and an image inputting means for inputting consecutive images in chronological order by the image input means. a changing area extracting means for extracting a changing area of a moving object using the changing area extracting means; a rectangle information acquiring means for acquiring rectangular information that includes the changing area extracted by the changing area extracting means; and a rectangle acquired by the rectangular information acquiring means. Rectangular information enlarging means for enlarging information into rectangular information having a predetermined magnification size; and image processing means for tracking a moving object by performing image processing on a corresponding area.

[0007]

[Operation] The moving object tracking device according to the present invention extracts and tracks moving objects using a two-step processing method. That is, first, a changing area of a moving object is extracted using inputted chronologically continuous images, and rectangular information that includes this extracted changing area is obtained. Next, this obtained rectangular information is expanded to rectangular information of a predetermined magnification, and based on this expanded rectangular information, image processing is performed on the corresponding area of the input chronologically continuous images. Track a moving object by doing this.

[0008] By doing so, it is possible to minimize the time delay in the result of extracting a changed region compared to the conventional method. Furthermore, it is possible to track a moving object in a manner that is less affected by the photographing environment of the image input means. Furthermore, when a moving object enters the monitoring area, it is always moving at an appropriate speed, and it is possible to sufficiently respond to movements once the moving object enters the monitoring area. That is, the moving object can always be captured even if the moving object comes to rest after entering the monitoring area. If this is considered as a detection process, there are advantages such as no need to consider the sampling time interval of the input image.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows the overall configuration of a moving object tracking device according to the present invention. That is, the image input unit 1 is mainly composed of, for example, a television camera, and performs processing such as A/D conversion on images from the television camera to generate digital multivalued images that are continuous over time. Enter as . A digital multivalued image input from the image input section 1 is stored in the image storage section 2 as digital information. The image storage unit 2 stores a plurality of chronologically continuous images in order to extract changing regions of a moving object, and these images are updated with new images as needed.

The changing area extracting unit 3 extracts the changing area of the moving object as a binarized image by performing a pixel-by-pixel difference calculation between a plurality of chronologically continuous images stored in the image storage unit 2. Extract. This binarized image of the change area is an image that is output as a continuous image in time series every time the input multivalued image is updated.

The rectangle information acquisition unit 4 acquires rectangle information that includes the changed area extracted by the changed area extraction unit 3. Here, the rectangle information is not the size of the rectangle adjacent to the changed area. In other words, the changed area is expanded to a rectangular size that does not touch the four sides of the rectangle. The parameters of this rectangular enlargement process are set so that the extracted change area of the input image captured at the next time will fit within the rectangle. Therefore, parameters such as the value of the sampling time of the input multivalued image or the moving speed of the moving object are taken into consideration. The above parameters become important values when tracking this moving object.

The rectangle information obtained by the rectangle information acquisition unit 4 is sent to the projection processing unit 5, which obtains vertical and horizontal projection information for the binarized image within the rectangle of the corresponding input image, and uses this projection information. By doing this, the changing area within the rectangle and the background area are separated. Furthermore, a process is added in which independent areas connected to the four sides of the rectangle are determined to be background areas and removed. The rectangular information obtained here is in contact with the changed area, and becomes the information on the changed area that is finally determined. The rectangle information for tracking processing, that is, the rectangle information enlarged by the rectangle information acquisition unit 4, is applied to the next captured image.

In this way, the changed area extracting unit 3 is used to obtain the first changed image for each intruding object, and once the rectangular information of the changed area is obtained, the above processing is performed based on this rectangular information. However, the changed area extraction unit 3 always performs change extraction processing due to new intrusion, and if there is a new intruding object, its rectangular information is added to the processing target as a tracking target area. .

The switching section 6 switches the processing of the changed area extraction section 3 and the rectangular information acquisition section 4 for one intruding object. That is, once a changed area is extracted by the changed area extraction unit 3 for a certain intruding object, the process is subsequently controlled by the rectangle information acquisition unit 4. This is performed every time a new intruding object is extracted.

The image display unit 7 displays input images,
When a changed area is extracted, the display mode of these input images and changed area portions is controlled, such as displaying the input image and the changed area portion in a superimposed manner.

Next, the flow of processing in such a configuration will be explained with reference to the flowchart shown in FIG. First, after inputting chronologically continuous images using the image input unit 1, a changing area of a moving object is extracted by performing subtraction processing between these images, and as a result, it is checked whether or not there is a changing area. However, if there is no change area, the image input process is repeated, and if there is a change area, a rectangle that touches the change area is found, and then this rectangle is enlarged n times. This is to set a larger rectangular area in the next image in order to track a moving object as described above.

After that, each vertical and horizontal projection information is obtained for the binarized image showing the corresponding change area within the obtained rectangle, and processing is performed to remove the area touching the four sides of the rectangle. Processing is performed to determine that an independent lump region within the object is a moving object. Through the above processing, the invading object has been extracted from the extracted change area. The tracking of this intruding object is
This is done by repeating the above process.

FIG. 3 illustrates an example of the present invention using an actual image. FIG. 3A shows a region of change extracted using a plurality of time-series images from the image input section 1. In FIG. Since the portion where the density difference between images is equal to or greater than the threshold value is output as a changed area, the area other than the actual intruding object is output in the entire image. In FIG. 4B, region A is extracted as a candidate based on the lumpiness of the changed portion, and this state is shown by a broken line rectangle. The figure (c) shows the result of extracting the lump region.
This shows a state in which changed portions H1 and H2 that are in contact with the four sides of the rectangle remain. If this is the area of the changing object, it will lead to incorrect recognition judgment of the moving object, so as a criterion for removing the changing areas H1 and H2, a judgment process is performed in which the changing area connected to the four sides of the rectangle is treated as the background. Do this. Figure (d) shows the results. Here, a rectangle A that is in contact with a block area that is not connected to the four sides of this rectangle is finally extracted as a changed area.

Furthermore, in order to track this changing area,
A rectangle N is created by enlarging the rectangle A by n times on the four peripheral sides. Tracking is performed by treating this rectangle N as a change area candidate for the next input image. In this way, by handling a rectangle enlarged by n times, even if the moving direction of the moving object is front, back, left, or right, the rectangle N
It will fit inside.

As described above, according to the above embodiment, after first extracting a change area candidate by subtraction processing between a plurality of consecutive images, the next input one is based on this extraction rectangle. A moving object existing within the rectangle is extracted from the images. By repeating this process, the moving object can be tracked, so that the time lag in the moving object detection results can be minimized compared to conventional methods.

[0022] Furthermore, when a moving object enters the monitoring area, it always moves at an appropriate speed, and it is possible to sufficiently respond to movements once the moving object enters the monitoring area. That is, the moving object can always be captured even if the moving object comes to rest after entering the monitoring area. If this is considered as a detection process, there are advantages such as no need to consider the sampling time interval of the input image.

Furthermore, conventionally, a plurality of successive images are used in the process of extracting a changed area, so if the photographing environment, for example, environmental elements such as wind or sunlight, change within the time to capture the plurality of images, , and as a result, many noise components were generated that caused erroneous recognition in the extracted image of the changed area, but in the above example, the changed area was extracted for one image captured at the minimum time interval. and tracking, the above problem can be solved.

It should be noted that the present invention is not limited to the above embodiments. For example, in the embodiment described above, a changed area is extracted by subtraction processing between a plurality of consecutive images. If it is determined that the area has changed, it may be possible to extract a changed area within a specified rectangle centered around the area where the change has occurred. In addition, in order to extract the cluster of changed regions that exists within the extraction rectangle after the changed region has been extracted,
In the embodiment described above, processing is performed to remove areas connected to the four sides of a rectangle from a binarized image. In addition, in order to take into account the case where the change area is continuous with the original change area, vertical and horizontal projection information is obtained for the binarized image within the rectangle, and this information is used to define the change area within the rectangle. can be taken out.

Furthermore, when a plurality of locations are extracted as candidates for changing regions, processing speed can be increased by assigning respective processing modules to these candidate rectangles.

[0026]

Effects of the Invention As described in detail above, according to the present invention, it is possible to minimize the time delay in the extraction result of a changing region, and even if a moving object comes to a standstill after entering the monitoring region, It is possible to provide a moving object tracking device that can extract a moving object and further minimize the influence of the photographing environment.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a moving object tracking device according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating the flow of processing in the same embodiment.

FIG. 3 is a diagram for explaining an example of the present invention using an actual image.

[Explanation of symbols]

1... Image input unit, 2... Image storage unit, 3... Change area extraction unit, 4... Rectangle information acquisition unit, 5... Projection processing unit, 6
...Switching unit, 7...Image display unit.

Claims (1)

[Claims] Claim 1: An image input means for inputting images continuously in a chronological order, and a changing region extraction method for extracting a changing region of a moving object using the chronologically continuous images inputted by the image input means. means, rectangle information acquisition means for acquiring rectangular information that includes the changed area extracted by the changed area extraction means, and enlarging the rectangular information acquired by the rectangular information acquisition means into rectangular information having a size of a predetermined magnification. Rectangular information enlarging means and rectangular information enlarged by the rectangular information enlarging means perform image processing on corresponding areas of the chronologically continuous images inputted by the image inputting means. A moving object tracking device comprising an image processing means for tracking an object.