JPS5925509B2 - Image tracking method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an automatic tracking method for a video image of a single target captured by a television camera, and more particularly to an image tracking method characterized by a video processing method using a scan compression memory as a medium. .

In the conventional image tracking method using the television system,
The system extracts signals from points in the photographed image that have particularly high brightness compared to the surrounding areas, internally generates a tracking signal according to the movement of the signal, and performs image tracking. It's summery.

However, this method requires that there be a point or part on the target that has particularly remarkable brightness, and that it always maintains a clear difference from its surroundings.

Depending on the brightness of the entire target, which changes from moment to moment, and especially the angle of the target, direct reflection of sunlight that occurs in a short period of time may occur on or around the target in the form of a pulse signal, or On the other hand, the brightness of the entire target decreases due to cloud formations, etc., and these must be avoided as they act as noise and cause a decrease in tracking performance. For this purpose, the target to be tracked is made to look extremely different from the surrounding background screen by applying special paint to create a contrast with the surroundings, or by lighting a lamp to increase the brightness of a certain part of the target. It was necessary to distinguish between Even if no treatment is applied to the target, sufficient tracking will not be possible unless the contrast ratio with the background is 30 to 50% higher in terms of the image voltage converted by the image pickup tube. It was difficult. Therefore, in the conventional bright spot tracking method, the distance between the television camera and the target is very short, and it is almost impossible to put it to practical use outdoors. The present invention eliminates these conventional drawbacks and provides an image tracking method that makes it possible to track a target at a relatively long distance.

In the image tracking method of the present invention, a target to be tracked is photographed by a television camera, the photographed target image signal is stored in one or more scan conversion memories, and is temporarily made into a still image plane. The image information that changes from moment to moment and is output by the television camera is compared with the still image information, and the image information is reduced or reduced to the image size most suitable for tracking regardless of the distance between the target and the television camera. By comparing the next image information that has been enlarged and similarly stored and processed with a time delay, and the image information that has been converted to the image size most suitable for the tracking, both images are The amount of deviation in the position of the information is obtained as a control signal, and the control signal is used to control a drive base for mechanically driving the optical axis direction of the television camera, so that the target can be tracked regardless of its shape or size. It is characterized by tracking.

An embodiment of the image tracking method according to the present invention will be described below with reference to the drawings.

---1 FIG. 1 is a block diagram showing the configuration of an image tracking system according to the present invention.

Now, the image of the target 20 photographed by the television camera 1 is input to the switching device 3 and is displayed at appropriate time intervals (for example, 1
(for each field) and output the images alternately to the scan conversion memories 4 and 6 as the first and second screens. The switching device 3 only needs to be able to electrically set the timing of the video output signal of the television camera 1 and supply it alternately to the scan conversion memories 4 and 6, so a general electronic switch circuit using a semiconductor or the like may be used. use,
It is possible to switch the flow of the television video signal in synchronization with the synchronization signal generated by the television camera 1. Because one scan conversion memory 4 or 6 cannot read and read signals at the same time due to its characteristics, in this embodiment two scan conversion memories 4 or 6 are prepared, and while one (4 or 6) is reading the target information, and the other (6 or 4) to read. The first or second image signal read from the scan conversion memory 4 or 6 is shaped by the waveform shaping circuit 8 or 9 as described later, and is temporarily stored in the storage device 10 or 11. If the target 20 is moving, the target information stored in the storage devices 10 and 11 will have a slight positional shift from each other, so these two pieces of information are compared by the comparing circuit 12. and detect the amount of positional deviation. The amount of deviation in this position is converted into an electrical phase as an error signal, converted into the same amount as the amount of deviation in the direction 180 degrees opposite to the direction in which the target is deviated, and sent to the drive base 2 of the TV camera 1. In addition, the optical axis of the television camera 1 is always directed toward the target 2.0 by a servo motor or the like (not shown) built into the stand 2.

This specific embodiment will be described in detail with reference to FIGS. 2 to 5. FIG. 2 shows an example of the screen state when the television camera 1 captures the target 20, and screens a and b were taken at a certain time interval from each other. It is. FIG. 3 shows waveforms showing the signal processing situation in the waveform shaping circuits 8 and 9 of FIG. First, television image information photographed by the television camera 1 as shown in screen a in FIG. 2 is stored in the scan conversion memory 4 shown in FIG. The signal path is switched, and television image information photographed with a slight delay in time from the state of screen a, that is, screen b, is stored in the scan conversion memory 6. and screen b
While the image is being read, the signal on screen a is extracted as a series of electrical signals as shown in FIG. 3a. To simplify the explanation, each extracted signal is shown as a representative of only one television scanning line. A skim level SL is set for each extracted waveform, and signals above or below a certain level are erased. Next, a pulse A synchronized with the television scanning signal is provided, and the video signal is waveform-shaped by a waveform shaping circuit 8 in accordance with this synchronizing pulse A to obtain a second waveform. Furthermore, the center position of the target 20 is detected from the waveform 2. Similarly, the image signal aperture on screen b in FIG. 2 is also waveform-shaped to obtain a shaped waveform, and the center position of the target 20 is detected. The detected target center positions of screens a} and b are stored in the next stage storage devices 10 and 11, respectively. The data stored in the storage devices 10 and 11 with a slight time lag are input to the comparator circuit 12 without being taken out at the same time. The comparison circuit 12 calculates the amount of movement of the target 20 using the target center position of screen a or b as a reference, and outputs it as a movement amount signal. The comparator circuit 12 has two positive and negative polarities so that the moving direction of the target 20 can be determined. An automatic gain control circuit (not shown) is used to compensate for fluctuations (strength and weakness) in the video human input signal level caused by the brightness and darkness of the target 20 photographed by the television camera 1 and the background containing it. . Therefore, if the changing video input signal is always kept constant and the skim level SL is set for the signal after gain control, it is possible to respond with a constant skim level, and the skim level can be adjusted according to fluctuations in the video signal. There is no need to change the SL. Further, the reflected light from the target 20 may appear dark or bright depending on the surrounding situation. For such level inversion of the target signal, in addition to the skim level SL described above, a skim level is provided with an appropriate distance from the skim level SL, and the background is always between these two levels. It is also controlled by an automatic gain control circuit. Therefore, the target 20 will be detected by one of the skim levels, and the signal extracted by one of the skim levels can be converted to one or the other using, for example, an inverting circuit (not shown). is converted into a signal and processed.
In the above explanation, only the case of the amount of movement in the horizontal direction was explained, but in order to track the target 20, it is necessary to simultaneously detect the amount of movement in the vertical direction. It is obtained by measuring the number of scanning lines and extracting the center of that number. On the other hand, goal 20
Or the size of the target image on the imaging surface of the television camera 1, which is caused by the movement of the television camera 1.
Ideally, it is desirable to store pixels over the entire captured screen due to changes in color and shape.

However, this requires storage elements on the order of 1.times.10@6 or more, which is difficult to implement from an economic standpoint. According to the present invention, a scale control device 13 is provided in order to eliminate such drawbacks. The scale control device 13 first measures and stores the scale of the target image photographed by the television camera 1, particularly the target image immediately before tracking is to be started, on the screen. Then, it detects the difference in size from the target that changes from the next moment, and continuously inputs a signal corresponding to the difference as a zoom control signal to the scan control circuits (not shown) of the scan conversion memories 4 and 6. . Therefore, the system automatically performs so-called electronic zooming, which extracts and reduces or enlarges the necessary target image from among the screens containing targets already stored in the scan conversion memories 4 and 6. be. The target image after the zooming process is taken out and subjected to signal processing in the waveform shaping circuit 8 or 9. Therefore, the image size output from the scan conversion memory 4 or 6 can be kept constant regardless of the distance between the target 20 and the television camera 1, and the image size can be kept constant regardless of the distance between the target 20 and the television camera 1. It can be set to any size that is most suitable for tracking. The details of the scale control will be explained below with reference to FIGS. 4 and 5. When starting tracking,
First, an electronic gateway is manually set for the tracking target 20 as shown in screen a in FIG. 4. The electronic gate gate is set, for example, by surrounding the leading edge of the target image contour with four electronic lines, and this surrounded rectangular frame is used as a gate signal. First, the video signal of the target 20 (FIG. 5A') is waveform-shaped to obtain a second shaped waveform, and the gate signal is made to match this waveform. The size of the gate signal can be arbitrarily obtained by the operator selecting the size of the target 20 as appropriate and adjusting the gate size. When the gate signal R and the target signal G match, the gate signal R at that moment is held using a hold circuit or the like (not shown). Next, when the target 20 moves and its size changes as shown in screen b' in Fig. 4, this image signal (Fig. Compared to the gate signal held at
The difference is extracted as a scale error signal. In this case, since the target is not necessarily large, the scan control circuit (not shown) of the scan conversion memory 4 or 6 is designed to be able to generate signals in both positive and negative directions based on the gate signal. input as a zoom control signal. Through the above operations, the scan conversion memory 4 or 6 always outputs a target image signal of a constant size. In the above description, a case has been described in which two scan conversion memories are used, but similar image processing is possible with one scan conversion memory.

An example is shown in FIG. In the same figure, image information of the target photographed by the television camera 30 is manually stored in the scan conversion memory 31. The image information read from the scan conversion memory 31 is waveform-shaped by the waveform shaping circuit 32 in the same manner as in the previous embodiment, and the center position of the target is detected. Waveform shaping circuit 32
The signal path is switched by the switching device 33, and the output of
It is stored in the storage device 34 or 35. The image information stored in the storage devices 34 and 35 with a slight lag in time is
At the same time, the signals are taken out and input to the comparator circuit 36. The comparison circuit 36 compares the respective target center positions that are slightly shifted in time, calculates the amount of movement of the target, and outputs the result as a movement amount signal. The following processing is the same as in the above embodiment. In the case of this embodiment, the time interval between the screens to be processed is slightly longer than in the above-described embodiment. Further, in these embodiments, a television camera mainly used in the visible light region was used, but if an infrared television camera, for example, is used instead, it can be used at night as well as during the day.

As an application example of the image tracking method according to the present invention, for example, 1
In conjunction with a measurement camera using 6 mm film, the movement of aircraft, automobiles, ships, etc., or the firing state of mouth targets, etc. can be accurately measured.

Furthermore, it is also possible to safely and accurately track and observe objects that require observation without contact and at a sufficiently safe distance, such as a vehicle transporting molten metal taken out of a blast furnace. As explained above, in the image tracking method according to the present invention, the scale control as described above is performed, so that from the entire screen photographed by the television camera 1, the very surrounding area including the target 20 necessary for tracking is It is possible to extract and process an image of only the target 20, and it is possible to completely eliminate similar signal sources (noise sources) occurring near the target 20.

Furthermore, since data processing for the target 20 is an operation within a specific gated range, the storage devices 10 and 1
1 may also be very small. Furthermore, since the target image is temporarily stored in a scan convergence unit and processed as a still image, it is difficult to avoid fluctuations caused by heat haze on the target screen being photographed or water vapor around the target, especially when tracking a long-distance target. It is possible to completely eliminate apparent changes in the position of the tracked target or blurring of the outline due to fluctuations in the target image.

[Brief explanation of the drawing]

FIG. 1 is a functional block diagram showing the image tracking method of the present invention, and FIG. 2 is an image of a target taken with a television camera, where screen a shows the initial position of the target and screen b shows the initial position. This shows the case where the target has moved slightly from . Fig. 3 is a diagram showing the waveform processing state in the waveform shaping circuit, and Fig. 4 is an image of the target taken with a television camera, where screen a' shows the initial position of the target, and screen b' shows the initial position. This shows a case where the target has moved slightly from its position. Fig. 5 is a diagram showing the waveform processing state in the scale control device, and Fig. 6 is a functional block diagram showing another embodiment of the image tracking system according to the present invention.1... TV camera, 2
... Drive stand, 3 ... Switching device, 4, 6
・・・・・・Scan Conversion Memo 18 8,9
...Waveform shaping circuit, 10, 11...Storage device, 12...Comparison circuit, 13...
Scale control device, 20...Target.

Claims (1)

[Claims] 1. In an image tracking method in which a target to be tracked is photographed by a television camera, and the target is detected and tracked from the image signal output from the television camera, first and second targets having different photographing times are detected from the output image signal of the television camera. at least one scan-comparison memory device for obtaining the second screen as a still screen; and an electronic gate for defining an area to be image processed for either the first or second still screen; a scale control device that controls the first and second still screens to the same scale by zooming up or zooming down the scale of the other image with reference to , and the electronically gated first or second still image; a waveform shaping circuit for waveform shaping the zoom-controlled second or first still image stored in the scan compression memory device to detect the center position of each target; and the detected center position of the target. and a comparison circuit that compares the stored center positions of the respective targets, detects the amount of deviation between the center positions of both targets, and outputs the detected amount of deviation between the center positions of both targets as a control signal, the control signal being the output of the comparison circuit. An image tracking method characterized in that the drive base of the television camera is controlled by the following means to track a target.