JPH0614743B2 - Image input type tracking device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an image input type tracking device for observing and tracking the position of a tracking target (object) such as a moving object or an intruder in an image. .

(Background Art) FIG. 4 is a diagram for explaining a conventional method for tracking an object in an image by using image processing. In the figure,
Images A and B are the temporally consecutive input images,
The same object has moved from the position of the image A to the position of the image B, and the hatched portion is the overlapping region of the images A and B. It is assumed that the object is at the position of the image A at the time t and is at the position of the image B at the time (t + 1).

The tracking of the object in such a case is performed by the following procedure. First, the object extracted from the input image is labeled at time t and a label is given.
Next, a label is given to the object extracted from the input image at time (t + 1). In addition, this label,
The logical product of the two objects marked with is calculated to obtain the overlapping part as shown by the shaded area. If the area of this overlapping portion is more than a certain extent, it can be seen that the label object and the label object are the same object.
The object can be traced by continuously repeating this process. According to this method, there is a problem that the labeling process for labeling and the process for measuring the area of the overlapping part are complicated and the processing time is long. Further, if dedicated hardware is used to shorten the processing time, there is a problem that the configuration becomes complicated and the cost becomes high.

(Object of the Invention) The present invention has been made in view of the above points,
The purpose is to use simple hardware
An object is to provide an image input type tracking device capable of tracking an object in an image.

DISCLOSURE OF THE INVENTION In an image input type tracking device according to the present invention, in order to achieve the above object, as shown in FIG. 1, an imaging unit 1 for obtaining an input image and a tracking target. When the input image does not exist, the reference image memory 2 that stores the input image in advance, the difference unit 3 that inputs the input image and the reference image at the same timing to obtain the absolute value difference, and the output of the difference unit 3 are set in advance. A counter that includes a binarizing unit 4 that binarizes with a threshold and a plurality of counters and that specifies the number of pixels for which the output of the binarizing unit 4 has a specific value (“1” or “0”). The pixel counter 6 that counts with the window counter, the window image memory 5 in which the image is divided into at least the same number of areas as the number of counters included in the pixel counter 6, and different values are stored for each area, and read from the window image memory 5. Value Counter selector 7 that specifies the counter by coding
And a position determining means 8 for determining the position of the tracking target by reading the value of each counter in the pixel counter 6 and obtaining the counter having the maximum count value. .

FIG. 1 is a block diagram showing the configuration of the present invention. The input image obtained by the image pickup means 1 such as a television camera is input to the difference means 3 at the same timing as the reference image in the reference image memory 2 holding the background image, and the absolute value difference between them is calculated for each pixel. To be done. As a result, the background is removed from the input image, and only the tracking target (object) is extracted as the difference image and input to the binarizing means 4 in the next stage.
In the binarizing means 4, the difference image is binarized with a preset value, and "1" is assigned to the object and "0" is assigned to the background portion. On the other hand, the window image memory 5 is also read at the same timing as the input image signal,
The data is input to the counter selector 7. The counter selector 7 decodes the binary data read from the window image memory 5, and outputs a select signal designating a counter to the pixel counter 6. For example, if the data for each pixel of the window image memory 5 is 6
If it is a bit, the number of select signal lines is 2 ⁶ = 64.

FIG. 2 is a diagram showing the relationship between the window image memory 5 and the pixel counter 6 in the present invention. The window image memory 5 previously holds a grid pattern, and in FIG. 2, holds 8 × 8 = 64 grid patterns. Data of 0 to 63 are sequentially stored in each square as 6-bit binary data.
The window image memory 5 is, for example, 32 pixels vertically × 4 pixels horizontally.
Assuming that it has a size of 0 pixel and can store 6-bit data for each pixel, each grid has 4 pixels in the vertical direction × 5 pixels in the horizontal direction.
Since each pixel has the size of pixel and 6-bit data can be stored for each pixel, all the 20 pixels belonging to one square store the same value. Pixel counter 6
Is provided with 64 counters so as to correspond to the squares of the window image memory 5 in a one-to-one correspondence. The read data from each pixel of the window image memory 5 is decoded by the counter selector 7, and the pixel counter 6
Of the 64 counters in, the operation of only one counter corresponding to the square to which the pixel belongs is permitted.
The counter permitted to operate counts the number of pixels of "1" in the corresponding grid in the input image. Since these counters operate in synchronization with the image signal, the number of pixels having the pixel value of "1" in one screen is divided into each square and counted in one field period (16.6 msec). The number of pixels having a pixel value of "1" in each square indicates which square area the object is in. Therefore, the count value of each counter after one screen measurement is set to the position. The position of the object can be found by searching the counter having the maximum count value read by the determination means 8. For example, in the position of the object as shown by the broken line in FIG. 3 (a), the maximum count value is "2".
It can be seen that it is located at the position of the grid of 0 ". By repeating this operation, the object can be tracked.

Even when there are a plurality of objects in one screen, it is possible to track the objects having the closest temporal positions between the screens as the same object. However,
In this case, the precondition is that multiple objects do not overlap each other.

If the area of one object is larger than the square and there are multiple squares with the maximum count value as shown by the broken line in FIG. 3 (b), the count value is Although software processing is required to average the positions of the plurality of cells having the maximum value, this calculation can be performed at high speed.

As is clear from the above description, when the size and shape of the tracking target on the image can be predicted, it is preferable to set the size and shape of the grid to be substantially the same as the size and shape of the tracking target.

(Effect of the Invention) As described above, the present invention has a simple hardware configuration by dividing an image into a plurality of watersheds and providing counters for counting the number of changed pixels in each area in correspondence with each area. However, it is possible to know the approximate position of the object in the image from the value of each counter, and because it can be configured by hardware, the position of the object can be known in real time, and high-speed tracking operation is possible. Has the advantage that

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of the present invention, FIG. 2 is an explanatory diagram showing the configurations of the window image memory and the pixel counter of the same, FIG. 3 is an operation explanatory diagram of the present invention, and FIG. 4 is a conventional example. FIG. 1 is an image pickup means, 2 is a reference image memory, 3 is a difference means, 4
Is a binarizing means, 5 is a window image memory, 6 is a pixel counter, 7 is a counter selector, and 8 is a position determining means.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Sada Sada, 1048 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Works, Ltd. (56) References JP 61-177090 (JP, A) JP 58-94271 ( JP, A) JP-A-56-31286 (JP, A)

Claims (2)

[Claims] 1. A pixel counter including a binary image input means for inputting a binary image to be tracked and a plurality of counters, wherein a designated counter counts the number of pixels having a specific value in the binary image. And a window image memory in which an image is divided into at least the same number of areas as the number of counters included in the pixel counter and different values are stored for each area,
A position that determines the position of the tracking target by decoding the value read from the window image memory and specifying the counter, and by reading the value of each counter in the pixel counter and finding the counter with the maximum count value. An image input type tracking device, comprising: a determination means. 2. The binary image input means includes an image pickup means for obtaining an input image, a reference image memory in which an input image when a tracking target does not exist is stored in advance, and the input image and the reference image are the same. The difference means for inputting at the timing to obtain the absolute value difference and the output of the difference means with the preset threshold value are set to 2
The image input type tracking device according to claim 1, characterized in that it comprises a binarizing means for binarizing.