DE2937284A1 - TELEVISION MOTION DETECTION - Google Patents

Description

ί Η: Sep.

1-0ICtI ₁ UiWoI!.:

Micro Consultants Limitui Dip. S ϋ ^ iamaw. Oipl. fh »j. Ur K. Fincke

«= _{Uee +} " ,, .. _β Oip :. ! r.3 Γ r. Weu'i-ann. CpI. Cfcem. B. !! about

5 West Mills .. ^ _x .ι,. ^; _t i _i5 i,

Newbury, Berkshire, England "f-wliiifraijs 22,? 0ü0 Mündren 85

TV picture motion detection

The invention relates to the detection of television picture movement, which is commonly used in various systems, for example for Noise reduction, image zooming and standards conversion is required.

In known motion detectors, for example, a single Using a pixel from two frames to determine if there has been movement can make it difficult for the detector be able to distinguish between motion and image noise.

The invention is based on the object of creating improved motion detection.

According to the invention, this object is achieved by a television image motion detection system solved, which is characterized by a detector device for detecting movement between a plurality of pixels and a controller for determining that movement has taken place when have moved at least some of the plurality of pixels in accordance with the detection by the detector device.

Furthermore, according to the invention, a method for image motion detection is provided specified in video information, which is characterized by the detection of movement between a plurality of pixels and determining that movement has occurred if movement is detected in at least some of the pixels became.

030013/0860

The invention is intended below with reference to the drawings, for example It shows:

Figure 1 shows a known noise suppression system, Figure 2 shows an alternative system,
FIG. 3 shows a simple motion detector, FIG. 1 shows an image area to be examined, FIG. 5 shows an arrangement for improved detection and FIG. 6 shows another exemplary embodiment of an improved detector.

As already mentioned, the motion detection is in various systems, such as noise reduction systems, image zoom systems and standards converters are required.

For example, it is in a noise reduction system in which part of the image information is added to incoming image information is added in order to effectively reduce the noise content in the picture, necessary when movement has taken place, the amount reduce the amount of information fed back when the moving part of the image is processed, otherwise smearing and other effects can be seen when viewing the video output on a television screen.

The basic mechanism for noise suppression is the digital integration of video data within a frame memory. An example of a noise suppression system is shown in Figure 1, as disclosed in greater detail in British Patent Application No. 12751/76.

030013/0860

The signal paths are digital signal paths. Each of the signal paths carrying the video data is capable of transmitting at least 8-bit wide video data with a typical clock frequency of 15 MHz.

New video data is received by subtracter 230, which also receives old data from memory 22. The subtracter output is received by the coefficient unit 231 for multiplying the subtracter output by a selected coefficient KI1. The output from unit 231 is added to the old video data in adder 232. The adder output is available for receipt by memory 22.

Channel 1 is an input channel to the digital frame memory which is capable of carrying a full grid of television information, each memory location being wider than 8 bits . In a practical system, a total of 12 bits can be used in any location in video memory. In practice , such a video memory would have a capacity of approximately 6 megabits. The video memory has at least one input channel and one output channel. It is more common for the video memory to have three channels as shown, one of which is an input channel while the other two are output channels. The arrangement of the memory allows the input and output to be asynchronous with respect to the television field and line frequencies . Channel 2 is able to run synchronously with channel 1, so that access to video data that is stored in the same picture position from a previous field can take place at the same time.

030013/0860

for which new information is available for the same pixel in the next field.

Thus, access to the top left pixel of field 1 from channel 2 is available at the same time as the top left Point of field 3 is available at the input of new video data.

A video memory of the type useful in such a system is disclosed, for example, in the UK patent application No. 6585/76 described in more detail.

An alternative noise suppression system is shown in FIG. While the subtracter 230 receives old and new data as before, the adder 232, on the other hand, receives new data at an input Data (and not old data) and, as before, receives the output of coefficient unit 231 at its other input.

In a simple noise suppression system, the values of the coefficient would be fixed, so that the parts of the previous data and the incoming data would be in a predetermined ratio to one another.

However, there is a conflict between the requirement for intoxication

suppression in which the highest integration time is used and the requirement that the picture should retain a representation of motion without distortion. For the latter - motion display the lowest integration time is required. Thus, it is desirable to provide an adaptive mechanism that is intelligent enough is the coefficients based on a variable integration time

030013/0860

to be adjusted depending on the image content. This leads to the need to create a motion detection.

If the incoming video information is immobile, for example a test image transmission - the coefficient kll for the purpose of optimal noise reduction can be set to a value so that a large proportion of the previous image information Pixel by pixel is added to the current information.

If the entire image is moving, for example while a camera is "panning", the coefficient kll must be set to one Another value can be set if the motion display is not distorted by the effect of the noise reduction system shall be.

If part of the image is stationary while a part continues to move, it is used to provide effective noise reduction to achieve compared to the stationary part of the picture and to enable a movement display without distortion, necessary. dig, kll based on pixel by pixel modify.

The first step to a more effective coefficient modification system, which adapts to the movement of the picture is the comparison of changes that occur between successive ones at the same place in the picture stored data values have occurred in continuous frames, and a motion detector is one such mechanism for determining that changes occur in the picture.

030013/0860

A simple motion detector is shown in Figure 3 and has a subtracter 205 which receives the old and new data and the output of which is passed to the comparator 206.

Each pixel in a new image is based on the previous one Place stored data is subtracted and the difference signal becomes applied to the comparator. When the difference signal exceeds a threshold level, the image is considered to be moving has, and the coefficient kll is then switched to a level that would be suitable for movement. As long as the change is within the threshold value, it becomes noise suppression coefficient proper of still images is used.

A read-only memory (ROM) could be provided instead of the comparator and the subtracter 205 is used similarly to the previous arrangement. The operation and structure of a Read-only memories are well known. The difference signal is now used to address the read-only memory, the different Contains coefficient values of kll in fixed memory locations. A distinction is made between large movements and small movements, and a sliding scale of coefficients on a real time basis is provided.

Motion detection, which uses subtraction between old video data and new video data, is on a single pixel basis has been described. However, it has been found that the difference between noise and real motion is more effective can be determined when examining an image area.

030013/0860

Figure k shows the area test. In the exemplary embodiment, 9 pixels are arranged in the form of a square with a single pixel in the middle. Line N has the pixels P ₁ to P-. In order to determine the total difference at image point P, -, a contribution is made from all surrounding image points. In one embodiment, this takes place in that each pixel on the surface is estimated with regard to the threshold value and that a logical gate circuit (e.g. a logical majority circuit) is used to detect the threshold value decision, so that it is signaled that a movement has taken place if it is determined that a portion exceeds the threshold.

In this system, four out of nine is a typical practical setting, to indicate movement. An example of such a system is shown in FIG. The difference signal will be like previously created by subtracter 205 and goes to comparator 206. The output goes to majority logic gate 250. There are only two of the nine gate inputs are shown in more detail, although each input has its associated subtractor and comparator.

Another method of area selection is not based on majority logic Made use of. In this method, the entire area is integrated so that the average difference is on the area can be compared with a threshold value. The aim is to differentiate between noise and movement. One such an arrangement is shown in FIG. 6 and comprises, as before, a subtracter 205 and a comparator 206, but in addition an integrator 300. The data from current and previous image

030013/0860

points for the given area are subtracted pixel by pixel, and the difference is integrated in the integrator 300. The integrator output is received by comparator 206 which effectively determines that movement has occurred when the integrator output exceeds a predetermined threshold value,

The system is particularly advantageous in a color television system using an NTSC or PAL color subcarrier. By Using the arrangement shown for surface movement detection, the system can be made insensitive to a residual auxiliary carrier will. The formation of a decoding circuit for eliminating residual subcarriers is simplified if a scheme is used, generally against the presence of residual sub-carrier components is insensitive.

Although the written area comprised 9 pixels, other numbers of one or more lines can be used. Ever the greater the number of integrated pixels, the closer the noise is averaged towards zero, with the movement against it remains undamped.

On the other hand, it is due to the occurrence of an intoxication "yard" around moving objects if too large an area is used. The intoxication "yard" comes from the size close to the linear dimensions of the surface in both the horizontal and vertical axes. Compared to this dimension it becomes impossible to tell the difference between noise and motion using area selection.

030013/0860

About the ability of the detection system to distinguish between noise »and movement to distinguish, to reinforce, without increasing the area on which the decision is made, is made by a combination made use of two fields. In this way it becomes possible to determine the number of points within the area (Volume) close to the pixel and thus dampen the noise without weakening the movement.

The outputs of Figures 5 and 6 could be used to control other arrangements, for example using of read-only memories, as described above, to vary the selected coefficients.

030013/0860

Blank page

Claims (10)

293728A Claims {I / television image motion detection system, characterized by detector means (205; 205, 300) for detecting movement between a plurality of pixels and control means (206; 206, 300) for determining that movement has taken place when at least some of the plurality of Have moved pixels due to the detection by the detector devices 2. Plant according to claim 1, characterized in that the detector device a subtracter (205) for sequentially detecting motion at each pixel within a selected part and that the control device has a comparator (206) for determining whether a predetermined threshold value has been exceeded at the output of the subtracter and a gate circuit (250) for detecting that the threshold value has been exceeded by a predetermined value Has number of pixels within the selected part. 3 · Plant according to claim 1, characterized in that the detector device a plurality of subtractors (205), each for detecting movement at a pixel in a selected Image part are determined, and that the control device has a comparator (206) for each of the subtractors for determination the exceeding of a preselected threshold and a gate circuit (250) for detecting the exceeding of the threshold value has a predetermined number of pixels within the selected part. 03001 3/0860 COPY 4. Plant according to claim 1, characterized in that the detector device at least one subtracter (205) for detecting the difference between pixels and an integrating device (300) for integrating the differences between each of the pixels within the selected part and that the control means a comparator (206) for determining whether a predetermined threshold value has been exceeded at the integrator output. 5. Plant according to one of claims 1 to 4, characterized in that that the plurality of pixels is selected from a part of a field and that the movement between pixels of a corresponding one Part of another field is detected. 6. Noise suppression system with a motion detector after a of claims 1 to 5, characterized in that the noise suppression system subtracting means (230) for subtracting incoming and previous image information, a modifying means (231) for modifying the output of the subtracter as a function of a selected coefficient depending on the detected one Movement and adding means (232) for adding the modified image information to the incoming or earlier image information having. 7. Method for capturing image movement on video information, characterized by detecting movement between a plurality of pixels and determining that movement has taken place when motion is detected at at least some of the pixels. 030013/0860 8. The method according to claim 7j, characterized by subtracting of pixels of corresponding parts of other fields for the purpose of motion detection. 9. The method according to claim 8, characterized by integrating of the differences of each of the subtracted pixels within the preselected image part. 10. The method according to claim 8, characterized by comparing any difference with a predetermined threshold to determine that movement has occurred if the threshold is exceeded at at least some of the pixels. 030013/0860