JPH04256078A - Moving body identification device - Google Patents

Abstract

PURPOSE:To always identify a moving body, by suppressing the occurrence of the noise of a difference value between picture elements owing to the change of illuminance in a moving body identification device detecting the moving body with a picture and identifying the moving body from the feature of the picture. CONSTITUTION:A background picture memory part 2 storing a background picture, a present picture memory part 3 storing a present picture, a difference operation part 4 operating a difference between the background picture in the background picture memory part and the present picture in the present picture memory with two thresholds, a comparison part 5 comparing the outputs of two difference pictures operated by two thresholds in the difference operation part 4 and a picture alteration part 6 rewriting the background picture in the background picture memory part by the signal of the comparison part 5 are provided. Thus, the occurrence of the noise of the difference value between the picture elements owing to the change of illuminance can be suppressed and only the moving body can always be identified.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for detecting a moving object in an image and identifying the moving object from the characteristics of the image.

0002

2. Description of the Related Art A conventional moving object identification device of this type uses a plurality of infrared sensors to detect the temperature of a moving object (for example, a human body), and detects the position from the direction of the infrared sensor that has a larger infrared output. There was a means to detect objects and moving objects using images. Means for detecting a moving object from an image include means for taking a difference screen C between the background screen A and the current screen B and extracting only the moving part, as shown in FIG.
There is a means to obtain the so-called image differential value, which is the difference F between the current screen E and the current screen E.

0003

[Problems to be Solved by the Invention] However, among the above-mentioned conventional methods, the means for detecting using infrared sensors requires complicated control because even though the approximate position can be determined, the accuracy is not high enough to recognize each part of a moving object. The problem was that it couldn't be done.

[0004] In addition, the method of taking a difference screen and extracting only the moving parts has the advantage that the shape of the moving object can be extracted because the background is removed, and the position of each part can be detected from the accurate shape of the moving object. However, there was a problem in that when there was a change in illuminance (for example, when the brightness in the room changed), noise was generated across the screen.

Furthermore, since the means for determining the differential value of the image is a differential value of the image that is obtained by calculating the difference from the previous screen, the problem of noise occurring on the entire screen is eliminated. but,(
1) Although the edge portion of the moving object is extracted, since it is not the entire moving object, it is difficult to detect the edge line when detecting the position because the edge line is broken, resulting in an incorrect position being detected. (2) When the movement of the moving object decreases, the edge line becomes thinner and eventually disappears. Therefore, there is a problem in that moving objects disappear or appear, causing malfunctions.

SUMMARY OF THE INVENTION The object of the present invention is to solve the above-mentioned problems, to suppress the generation of noise in the difference values between pixels due to changes in illuminance, and to make it possible to always identify only moving objects.

[0007]

[Means for Solving the Problems] In order to achieve the above object, the present invention provides a background image memory section for storing a background image, a current image memory section for storing a current image, and a background image memory section in the background image memory section. and a difference calculation unit that performs a difference calculation between the current image and the current image in the current image memory using two threshold values, and a comparison unit that compares the outputs of the two difference images calculated using the two threshold values in the difference calculation unit. The apparatus further includes an image changing section that rewrites the background image in the background image memory section based on a signal from the comparing section.

[0008]

[Operation] With the above configuration, the present invention calculates the difference between the background image in the background image memory section and the current image in the current image memory section using two threshold values, and calculates the difference between the high threshold difference image and the low threshold difference image. The comparison section compares the difference with the difference image, and the background image is corrected by rewriting the background image based on the signal from the comparison section.

[0009]

Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings.

In FIG. 1, 1 is an imaging means, 2 is a background image memory section for storing the background image input from the imaging means 1, 3 is a current image memory section for storing the current image input from the imaging means 1, Reference numeral 4 denotes a difference calculation unit that calculates a difference between the background image in the background image memory unit 2 and the current image in the current image memory unit 3 using two threshold values; A comparison section 6 compares the outputs of the difference images; 6 an image change section that rewrites the background image at the location of the image that appears as a difference in the comparison section 5; and a control section that controls equipment (for example, an air conditioner) that acts on the obtained moving object. 8 is a CPU, 9 is a bus, and 10 is a moving object. In FIG. 2, 11 is a background image, 12 is a current image, 13 is a low threshold difference image, 14 is a high threshold difference image, 15
is a threshold difference image between the low threshold difference image 13 and the high threshold difference image 14, the subscript a is the initial image, and the current image is changed in the order of b, c, and d. Note that 16 indicates a window, and 17 indicates light incident from the window 16. Subscripts e, f, g, h and i
represent a moving object 10, a window 16, and a light 17 in a background image 11, a current image 12, a low threshold difference image 13, a high threshold difference image 14, and a threshold difference image 15, respectively.

The operation of the above configuration will be explained based on the image schematic diagram of FIG. A background image is input by the imaging means 1, and the background image 11a is stored in the background image memory section 2. Next, an image is inputted by the imaging means 1 and stored in the current image memory section 3 (the image stored here is referred to as a current image 12a). After that, the difference calculation section 4 calculates the difference between the current image 12a and the background image 11a (hereinafter, the term difference calculation also refers to an image that has been subjected to binarization processing etc. after calculation).
A difference calculation is performed using a predetermined threshold value set in advance. Here, two types of difference calculations, a low threshold value and a high threshold value, are performed to obtain a low threshold difference image 13a and a high threshold difference image 14a, respectively. In the low-threshold difference image 13a and the high-threshold difference image 14a, the windows 16e and 16f cancel each other out, and since the moving object 10 does not exist in the current image 12, no result of the difference calculation is extracted. After that, when the moving object 10 appears in the current image 12b, the difference calculation unit 4 calculates a difference image between the current image 12b and the background image 11a, and each low threshold difference image 1
3b, obtain a high threshold difference image 14b. Low threshold difference image 1
3b, the high threshold difference image 14b has the motion 10f in the current image, so extracted moving objects 10g and 10h are obtained. Here, the comparison section 5 compares the image outputs of the low threshold difference image 13b and the high threshold difference image 14b, but in this case, since both images are the same, there is no difference, so proceed to the next step.

Next, when light 17f enters the current image 12c from the window 16f, the current image 12c and the background image 11a
When the difference calculation is performed by the difference calculation unit 4, a low threshold difference image 13c and a high threshold difference image 14c are obtained, respectively. Since the moving object 10f exists in the current image, the low threshold difference image 13c and the high threshold difference image 14c obtain the extracted moving objects 10g and 10h, but the low threshold difference image 13c has a difference of more than the threshold from the background image 11a due to the presence of the light 17f. Therefore, the extracted light 1
7g is extracted, and the high threshold difference image 14c contains the light 17g.
f is absorbed within the threshold and there is no extracted light 17. After that, in the comparison unit 5, the low threshold difference image 13 and the high threshold difference image 1 are
The threshold difference image 15c is obtained by performing the comparison of 4. In this threshold difference image 15c, only the extracted light 17g appears as a difference,
A variable light 17i is obtained. In this way, when a changing portion exists in the threshold image 15 (here, the changing light 17i), the image changing unit 6 rewrites the image data at the location where the changing light 17i exists at the same location in the background image 11.

Next, using the background image 10d as a background, the difference calculation section 4 performs a difference calculation between the current image 12d and the background image 10d. And a low threshold difference image 13d and a high threshold difference image 14, respectively.
get d. Here, both the low threshold difference image 13d and the high threshold difference image 14d are windows 16e and 16f, lights 17e and 17f.
The difference between them is canceled out, and only the moving object 10f in the current image is extracted as the moving object 1.
Obtained as 0g, 10h. Thereafter, the comparison unit 5 compares the low threshold difference image 13 and the high threshold difference image 14 to obtain a threshold difference image 15d. Since nothing is extracted from this threshold difference image 15d, the background is not rewritten by the image converter 6. Note that the low threshold difference image 13 and the high threshold difference image 1
4 and the threshold difference image 15 are binarized.

The above operations are controlled by the CPU 8 via the bus 9, and the control section 7 controls the devices that act on the moving object obtained by the difference calculation section 4, comparison section 5, and image modification section 6. . Here, moving objects are extracted using low threshold difference images,
If there is a sudden large change like the low threshold difference image 13c, the extraction may be omitted.

In the case shown in FIG. 2, the high-threshold difference image is exactly the same as the low-threshold difference image (excluding the low-threshold difference image 13c), so it is better to use the high-threshold difference image for moving object extraction. However, since a high-threshold difference image may make the image thinner, a low-threshold difference image is usually used to extract the moving object.

According to the configuration of this embodiment, the difference calculation unit 4 calculates the difference between the background image in the background image memory unit 2 and the current image in the current image memory unit 3 using two threshold values;
A comparison unit 5 compares the difference between the high threshold difference image and the low threshold difference image that have been subjected to the difference calculation, and the signal of the current image at the location of the image that appears as a difference in the comparison unit 5 is changed to the background image by the image change unit 6. Rewriting to the same location has the following effects: (1) illuminance can be corrected, and (2) moving body parts can always be extracted.

[0017] Instead of using a camera as the imaging means, images may be directly inputted into the background image memory section from a recording/playback device that has previously recorded images and used for analyzing the movement of a moving object.

[0018]

[Effects of the Invention] As explained above, the moving object identification device of the present invention calculates the difference between the background image and the current image using two threshold values, and compares the difference between the images subjected to the difference calculation in the comparing section. By rewriting the background image that appears as a difference in the comparison section, the following effects can be obtained.

(1) Since the moving object is extracted using a difference image between the background image when no moving object exists and the current image when the moving object exists, the shape of the moving object can be extracted accurately.

(2) If there is a change in the background image, such as when light is incident on the background, images calculated using thresholds of different levels are compared, and if there is a change, the background is rewritten. Illuminance can be corrected.

[Brief explanation of the drawing]

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

FIG. 2 is also a schematic diagram of the operation.

FIG. 3 is a schematic diagram of the operation of a moving object identification device in a conventional example.

FIG. 4 is also a schematic diagram of the operation.

[Explanation of symbols]

2 Background image memory section 3 Current image memory section 4 Difference calculation section 5 Comparison section 6 Image change section

Claims (1)

[Claims] 1. A background image memory unit that stores a background image;
a current image memory unit that stores a current image; a difference calculation unit that calculates a difference between the background image in the background image memory unit and the current image in the current image memory using two threshold values; A moving object identification device comprising: a comparison section that compares outputs of two difference images calculated using two threshold values; and an image change section that rewrites a background image in the background image memory section based on a signal from the comparison section.