JPH01282983A - Method and device for monitoring by tv camera - Google Patents

Abstract

PURPOSE:To standardize the position of a moving object with high accuracy in a moment by calculating the position according to a triangulation method after detecting the moving object by comparing input information with the preceding information and expressing the image of the moving object with dots. CONSTITUTION:Image information from infrared TV cameras 5 and 6 are alternatively sent through a communication equipment to an image input and output device and hourly inputted to a picture memory as digital information. The input information is compared with the preceding information, and a monitored object is detected as the moving object when difference information equal to a prescribed area or more is obtained. The image of the moving object is expressed with the dots, and the position of the moving object expressed with the dots is calculated according to the triangulation method.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides a monitoring method for detecting when an abnormal moving object enters a monitoring area using two or more TV cameras; The above also relates to the monitoring device. The above-mentioned abnormal moving object means that it does not normally exist in a steady state. Therefore, an abnormal moving object is not necessarily an object associated with a failure.

The abnormal moving objects to be monitored by the present invention include, for example, smoking and nine people leaking water.

[Conventional technology]

This is a conventional technique that automatically detects the coordinates of an object by using two television cameras and processing the captured image information.

(i) As disclosed in Japanese Unexamined Patent Publication No. 58-70108, the surface of the object is irradiated with a spotlight placed in the vicinity of the TV camera, and the surface is scanned automatically or manually along the shape of the object, A technology that processes the reflected light using two TV cameras to measure the position and shape of an object.

(ii) As disclosed in Japanese Patent Application Laid-Open No. 60-140113, image processing is performed on an object by making the optical axes of two TV cameras parallel, and by superimposing each image, the amount of shift is used to calculate the amount of the object using the principle of triangulation. Using technology to measure the coordinates of an object and (City) two TV cameras, the object is extracted by pattern matching, etc., the center coordinates are determined, and the three-dimensional object is calculated by triangulation from the azimuth angle of each TV camera. Techniques for measuring position are known.

[Problem to be solved by the invention] Among the known techniques (i), (ii), and (ni) above,
(i) In the method using a spot, (b) It can only be used in areas where the spot light hits, and the TV
It is not possible to extract a moving object that appears at any position within the field of view.

Further, it takes time to scan the spot light to find the center of the object.

0) If the object is uneven, measurement cannot be performed because reflected light cannot be obtained.

(c) A spot light generator is required, resulting in a large size.

(n) In the method of image processing an object using two TV cameras with their optical axes parallel, the distance between the TV cameras is
If the TVs are spaced 2 times apart vertically, horizontally, or horizontally, or if the optical axes are not parallel, the shape, size, etc. of the images captured by each TV will be different, and simply superimposing the images will not be accurate, so it is difficult to measure accurately. It takes a huge amount of time.

(■) Extracting a moving object from the background of an object outdoors, including the method (ii) above, involves many disturbances and requires an enormous amount of processing time.

Due to these problems, it could only be used under certain conditions.

The present invention was made in view of the above-mentioned circumstances, and does not require spotting equipment, and can detect a moving object that has entered the field of view of a TV camera and instantly calculate its coordinates with high precision. It is an object of the present invention to provide a monitoring method capable of making measurements and a device suitable for carrying out the above method.

Means for Solving Problem C8] Regarding the present invention created to achieve the above object,
The basic principle will be briefly explained below.

Either a conventional visible light TV camera or an infrared TV camera can be used to practice the present invention.

For example, if we consider the case of monitoring the appearance of a moving object such as a person or smoke (an object that is not normally present in a steady state), the person or smoke can be detected by either a visible light camera or an infrared camera.

And the present invention can be applied to such people and smoke (this is just one example).
The information representing the state when the moving object does not exist is stored in advance6. Then, the information before the appearance of the moving object is stored moment by moment as a binarized or multivalued image processing video, and the difference from the previous image is stored. Perform calculations.

When a moving object appears within the field of view, a difference image appears based on the difference in the amount of light or heat of the moving object with respect to the information before its appearance. Therefore, if it exceeds the electric image area and the output voltage corresponding to the amount of light or heat is higher or lower than a certain level, it is determined as a moving object, and this level is also used as a threshold for binarization. do.

Next, the difference image having a certain area is reduced evenly by a certain number of pixels in the X and Y directions of the screen. Finally, the above reduction process is stopped leaving only about one pixel or a constant value for each of X and Y, and the center of the area at this time is taken as the center position of the moving body.

Next, the azimuth angle from the center of the image to the center of the moving body is automatically calculated.

The above processing is performed for each of the two TV cameras, and the position of the moving body is automatically calculated based on the principle of triangulation.

[Effect]

An object seen on a TV camera or an infrared TV camera is extracted from the background based on the difference between the object and the background. Here, the shape of the moving object is a two-dimensional shape with a certain area, and there are some microscopic particles around it that are caused by differences in brightness or heat levels caused by state changes with time changes in the background before and after the appearance of the moving object. The difference image remains as noise.

From this state, all images on the image are reduced equally in the horizontal and vertical directions. At this stage, peripheral minute noises are removed, and the remaining moving object image ultimately becomes a dot or line image at a one-pixel level. When the reduction process is performed by one pixel each in X and Y and stopped when a bell remains, the moving object becomes dot-like. In this process, the image of the moving object is reduced evenly from the outer periphery in both the horizontal and vertical directions, so the coordinates of approximately the center of the image area remain, and as a result, the center coordinates of the moving object are calculated. Ru. Furthermore, all of the above is automatically processed based on the information stored in the image memory, so instantaneous measurement is possible. Therefore, moving objects appearing within the field of view can be reliably extracted and their positions calculated.

〔Example〕

Next, one embodiment of the present invention will be described with reference to FIGS. 1 to 4.

In this example, an infrared TV camera is used to monitor a moving object that has a higher temperature than a background object. In the present invention, monitoring refers to the work of discovering and locating objects that, in principle, do not exist during normal operation.

FIG. 1 is a detailed explanatory diagram of the present invention, and FIG. 2 is an explanatory diagram of the general arrangement of the apparatus of this example.

As shown in FIG. 2, a remote point 7 that is a blind spot (an area that cannot be seen directly) separated by an obstacle 8 from a moving object is set as a monitoring base. Two infrared TV cameras 5. 6 will be installed.

The coordinates of the TV left camera are (X++ yI+ z+), and the coordinates of the TV left camera are (Xzy y2+ zz).

When implementing the present invention, it is also possible to install three or more TV cameras.

It was captured by two TV left cameras, 6. The image based on the distribution of the infrared radiation intensity is in a multi-level gradation state as shown in 1 in Figure 1 (A-t), and as it is, it is impossible to determine whether there is a moving object in the screen or not. One feature of the present invention is that even if the threshold value is used, it is not known whether a moving object is present or not. Therefore, one feature of the present invention is that the appearance of a moving object is determined based on the state before the moving object appears. In other words, it is considered that a moving object (for example, a person or smoke) has some kind of heat, and is usually higher in temperature than the background. (
As will be detailed later, the present invention can be applied even at lower temperatures than the background.)

When a moving object enters the monitoring area (field of view of the TV camera), the peak of the infrared intensity in the original background is used as a threshold, as shown in Figure 1 (A-2), and the image is detected at regular intervals. Compare with new image. When a moving object appears, the difference from the threshold value is output as a difference in multivalued images, as shown in 2 in FIG. 1 (B-t). The figure (Ilz) is a chart showing the difference from the threshold value, and 4' is the binarized background noise.

Therefore, the moving object can be completely extracted by uniformly displaying the positive output part as a constant voltage as shown in 3 in Figure 1 ((,1), and binarizing the other parts as 0. This image can be easily converted into dots by horizontal and vertical reduction operations, etc., and the center position of the moving body on the image nχv ny
can be automatically calculated.

Also, simply about all pixels based on the background image.

A similar function can be obtained by calculating the difference from the coordinates before the appearance of the moving body.

Furthermore, the difference method has the advantage that even if the temperature of the moving object is lower than the background, the moving object can be detected as a difference.

In this case, the absolute value of the difference is taken and the same identification as above is performed.

4' is a reduced background nozzle. Figure 2 ((,2)
shows the signal waveform in (C-t) above, W is the signal width (time width) of the reduced image, and this ((,
2) What is indicated by the chain line in the figure.

It represents the signal before reduction.

Next, the camera angle of view is determined from the result of the mobile object identification operation described above. Figure 1 (Reduced image 4 of the moving object shown in (,1)
From this, its central coordinates nχv nF can be found. From this, the horizontal angle θ2 = 2・ny and the vertical angle 0 times = violet・ny (where klpk2 is a constant determined by the camera). Therefore, using this angle of view with the two TV cameras, Find the tertiary coordinates of the moving object as seen from. 2
The reference positions of the TV cameras are respectively o and y OZm.
If the reference direction is set to one direction, using the angle of view information, Ω?
Ll o 嘗, θ2 can be calculated.

As1nal +Bs1nα2=x4- xI--(1
)Acosc5+Bcosaz=yz yl---・
-C2) That is, the horizontal angle, elevation angle, or their absolute coordinates as seen from Oo can be calculated. Of course, by looking at the coordinates of two points on the moving object at a fixed time interval, the speed of the moving object can also be calculated. Further, a specific example of a system for performing the above calculation is shown in FIG.

Video information from the infrared TV cameras 5 and 6 is alternately sent to the video input/output device via the communication device. The image that has been digitally processed as a multi-tone image is input to a video memory and is stored in the form of the multi-level image 1 described with reference to FIG.

Therefore, this multivalued image is subjected to the above-mentioned moving object extraction calculation by the preprocessing circuit 15, and 01 and 02 are calculated here. Next, based on the coordinates and directions of the cameras 5 and 6 stored in the program memory 17, the program
and 02, the three-dimensional coordinates of the moving body viewed from point 00 are calculated. These series of operations, from extraction to calculation, are all instantaneously and automatically performed.

According to the above embodiment, a wide area can be monitored using the images of two TV cameras without any special restrictions on the placement of TV cameras, and without the need to focus on the spot, and when a moving object appears, it can be immediately and automatically monitored. can be detected and located.

〔Effect of the invention〕

By using the monitoring method of the present invention using the monitoring device of the present invention, there is no need to install spot equipment, and moreover, a moving object that has entered the field of view of the TV camera can be detected and its coordinates can be located instantly with high precision. It has excellent practical effects.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram of an embodiment of the monitoring method according to the present invention, FIG. 2 is a schematic layout diagram of an embodiment of the monitoring device according to the present invention, and FIG. 3 is an explanatory diagram of an embodiment of the monitoring method according to the present invention. FIG. 4 is a block diagram of the monitoring device according to the embodiment. 1...Multi-value image, 2...Moving object multi-value image, 3...
Moving object binary image, 4... Moving object reduced image, 5, 6...
・TV camera, 7... Remote point, 8... Obstacle, 9
...Communication device, 10...Switcher, 11...TV camera controller, 12...Image input/output device, 13.
... Image memory, 14... TV monitor, 15... Preprocessing circuit, 16... Output device, 17... Program memory, 18... Processing unit. 19... Image processing device, 20... Distance sensor. Representative Patent Attorney Tadashi Akimoto Jitsu Z Diagram 3rd Ward

Claims (1)

[Scope of Claims] A monitoring method that captures images of a moving object using one or more TV cameras and calculates the position of the object through image processing, comprising: (a) two TV cameras with known spatial coordinates and directions; (b) Compare the above input information with the previous information and periodically calculate the difference; (c) When differential information of a range level and a predetermined area or more is obtained, this is detected as a moving object, and (d) the entire screen of the TV camera is reduced to create an image of the moving object. A method for monitoring a moving object using a TV camera, characterized in that: (e) calculating the position of the point-formed moving object by a triangulation method. 2. The imaging range of the two or more TV cameras is defined as a monitoring area, and a calculation means is provided in a blind spot separated by an obstacle with respect to the monitoring area,
The TV according to claim 1, characterized in that camera image information is input to the calculation means through communication means.
Surveillance method using cameras. 3. Based on the image information captured by the two or more TV cameras and the positional information regarding the detected object calculated by the calculation means provided at the blind spot position, an image similar to that obtained when viewing the surveillance area directly from the blind spot position. A monitoring method using a TV camera according to claim 1, characterized in that the information is created by calculation and output to a monitor. 4. The difference calculation in item (b) above is a sweep between a steady state level stored in advance for each pixel and the level after the moving object is detected, as either the level of image brightness or thermal information. The monitoring method using a TV camera according to claim 1, characterized in that the monitoring method is carried out by calculation. 5. The difference calculation in item (b) above is performed based on the level of either the image brightness or thermal information, which is stored in advance for each pixel at a steady state and the peak after the moving object is detected for all pixels. 2. The monitoring method using a TV camera according to claim 1, wherein the monitoring method is carried out by a sweep calculation with a level aligned to a point. 6. The monitoring method using a TV camera according to claim 1, wherein the TV camera is an infrared camera. 7. The monitoring method using a TV camera according to claim 1, wherein the TV camera is a visible light camera. 8. The scope of claims characterized in that the reduction and dotization process in item (d) is characterized in that the image is reduced from the outermost periphery of pixels until an image of one pixel remains in each of the X and Y directions on the screen. The monitoring method using the TV camera described in 1. 9. In a monitoring device that captures images of a moving object using two or more TV cameras and calculates the position of the object through image processing, (a) Image information from two TV cameras whose spatial coordinates and direction are known. (b) an arithmetic unit that compares the above input information with the previous information and regularly performs a differential operation; When difference information of a level and a predetermined area or more is obtained, this is detected as a moving object, and the entire screen of the TV camera is reduced to make the moving object image into dots. A monitoring device using a TV camera, characterized in that it is provided with a calculation means; and (d) a calculation unit that calculates the position of the point-shaped moving object by triangulation. 10. The TV camera monitoring device according to claim 9, wherein the TV camera is a visible light TV camera. 11. The TV camera monitoring device according to claim 9, wherein the TV camera is an infrared TV camera.