GB1605019A - Video signal processor - Google Patents

Description

(54) VIDEO SIGNAL PROCESSOR (71) We MULLARD LIMITED, of Abacus House, 33 Gutter Lane, London, E.C.2. A British Company, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: This invention relates to apparatus for processing video signals. More particularly it relates to processing apparatus adapted to provide an indication of the presence of a relatively sharply defined and contrasted object in a scene in spite of electrical noise in the video signal and in spite of the presence of other objects of comparable contrast in the scene, which might otherwise give undesired or false indications.

The invention provides apparatus for processing a video signal comprising means for sampling the instantaneous value of the video signal at regular intervals of time, means for temporarily storing each sampled value, means for obtaining a differential signal as the difference between each instantaneous value of the video signal and the preceding stored sampled value, means for forming a running average of said differential signal related to a predetermined preceding time interval of said video signal, means for forming a running threshold level from the said running average of said differential signal, and means for providing an object indicator signal when the differential signal exceeds the running threshold level.

The video signals, consisting of a sequence of picture element signals, may be derived from a scene by using a conventional television camera to view the scene or by scanning the scene with a detector responsive to a region of the electromagnetic spectrum in which the scene is radiant. For example, the region may be the visible region, the near infra-red region or the far infra-red region otherwise known as the thermal infra-red region. The scene may be radiant as the result of natural or artificial illumination or, in the case of the thermal infra-red region, as a result of the temperature of the scene detail.

The video signals may be a continuous video waveform which may be regarded as a sequence of picture element signals following one another at a rate determined by the bandwidth of the video signal. The samples of the video signal may be taken at the picture element rate or a multiple of it.

Various forms of average of the differential signal may be used. For example, the averaging circuit may respond to the mean positive value of the differential signal.

Alternatively the root mean square value of the differential signal may be used. The time constant of the averaging circuit and hence the number of preceding difference signals which are used to form the average is a design parameter of the system. This time constant will generally be large compared to the duration of one picture element.

The compound of the average difference signal with the reference level to form the threshold level may be such that the average difference signal is dominant. Alternatively, the threshold level may be merely modified by the average difference signal.

If the scene being viewed or scanned contains large areas in each of which the scene radiance changes slowly as a function of distance across the area, the difference signals will be almost constantly small and hence the threshold level will be low. In this event a relatively small change of radiance from one picture element to an adjacent picture element will be sufficient to exceed the threshold and an object indication will be given. However, where scenes having a high content of sharp edged and/or highly contrasted objects are viewed, the difference signals will be larger, on average, and the threshold level will be raised. Object indications will only be given in this case where there is a relatively large change of radiance from one picture element to another. The rate of false object indications will be reduced in this second case over the rate which would have been obtained if the threshold level had remained at the value obtained with the scene of generally low contrast.

The indication may be fed to a visual or audible alarm circuit. It may also be fed to a display device showing the location of the object in relation to the scene. For example, if a television camera is used to view the scene, the indication may be superimposed as a characteristic signal on the video waveform fed from the camera to a display monitor. If the scene is scanned by a detector, a synchronous image reconstruction device may be provided to which the indication, alone or added to the detector output, may be fed. The object indicator signal may be of binary form, having one binary value when the differential signal exceeds said running threshold level and the other binary value when the differential signal equals or is less than said running threshold level.

The predetermined preceding time interval of said video signal over which the running average differential signal is formed may be chosen to suit the application. The interval may include only a few immediately preceding picture elements in which event the average changes during a typical video raster line which may contain up to a few hundred picture elements. The time interval may be as long as the duration of one such raster line and hence the average will not change along a line but only during the longer frame scanning time. The interval may be comparable to an entire frame time so that the threshold is the same for a whole frame, but changes from frame to frame.

An embodiment of the invention will now be described, by way of example, with reference to the accompanying drawing which shows a signal processing circuit, in block diagram form, adapted to process positive-going differences only.

Referring to the drawing the circuit comprises a sample-and-hold circuit 1, a differential amplifier 2, an averaging circuit 3, a comparator 4 and resistors R1 and R2.

The sample-and-hold circuit 1 receives the continuous video waveform input applied to terminal 5. Sampling pulses at the picture element repetition rate are applied to terminal 6. In between sampling pulses a stored value of the video waveform is available continuously at output 7. At each sampling pulse, the previously stored value is erased and the value of the video input available at the instant of the sampling pulse is stored. Typically, the sample-and-hold circuit 1 will contain a capacitor which is connected to the video input 5 by the sampling switch and which charges rapidly to the instantaneous value of the video input. At the conclusion of the sampling pulse the capacitor is isolated from the input 5 and the voltage on the capacitor is made available to output 7 via a high input impedance isolating device having unity gain. The differential amplifier 2, has a single output terminal 8 and two inputs 9 and 10 of opposite polarity. The gain of the amplifier from either terminal 9 or 10 to the output 8 is the same and may be low, tically unity. The output 8 then gives the difference of two positive voltages applied at the inputs 9 and 10 respectively. The positive polarity input 10 is fed directly from the video input 5 and the negative polarity input 9 is fed from the output 7 of the sample-and-hold circuit 1. Thus, immediately following a sampling pulse, the two inputs of the differential amplifier are the same since there has been no time for the video input 5 to change and hence the output at terminal 8 is zero. However, as the time elapsed since the sampling pulse increases, the video input will change, in general. The difference amplifier output 8 will depart from zero in either direction, depending on the sign of the video input change. At the conclusion of the sampling period, just before the next sampling pulse, the output 8 will have reached a value corresonding to the difference between two picture elements. During the sampling period, therefore, a roughly triangular pulse is generated at output 8 on the assumption that the video input 5 deviates from the stored output 7.

At the next sampling pulse, the output 8 will again fall to zero. Succeeding sampling pulses thus give rise to a succession of such pulses, some positive and some negative.

These difference pulses are applied to the input 11 of the averaging circuit 3. The output 12 is a quasi - d.c. signal from which the pulse nature of the input 11 is entirely removed and which reflects an average value of a predetermined number of preceding difference pulses. The output 12 may be the mean value of the positive-going pulses only. Alternatively, the pulse input may be treated as an a.c. signal and the root mean square value of a number of pulses evaluated to supply the output 12. The averaging circuit 3 may contain pulse averaging or a.c.

rectification circuits either of which may employ smoothing circuits having a time constant which will generally be long compared to the duration of one picture element. Such a time constant will effectively predetermine the number of preceding difference pulses which are taken into account in forming the average difference signal.

Such a time constant will give a decaying weight to earlier difference signals. Other forms of averaging circuit may be used in which difference signals earlier than a predetermined time are rejected altogether in forming the average difference signal.

The output 12 of averaging circuit 3 is compounded by resistors R1 and R2 with a reference level voltage applied to terminal 13. The ratio of these resistors is chosen to provide a desired threshold level voltage at the junction of the resistors which is applied to the negative polarity input terminal 15 of comparator 4. Comparator 4 is a high gain differential input d.c. amplifier in which the output on conductor 14 is limited not to fall below zero volts and not to rise above a fixed positive voltage V. These two values 0 and V are taken in further processing and/or display circuits to represent the binary digits '0' and '1' respectively. If at any time therefore, the difference signal at terminal 8 is more positive than the threshold level voltage by a very small amount inversely proportional to the comparator gain, the output at terminal 14 will be driven to the upper limit V and a '1' digit will then be allocated to the corresponding picture element. If the difference signal is equal to or more negative than the threshold level voltage, the output at terminal 14 will be driven to the low limit zero and a '0' digit will be allocated to the corresponding picture element. The ratio of the resistors R1 and R2 and the reference level voltage applied to terminal 13 may be chosen so that threshold level voltage is dominated by the output level of the averaging circuit 3 or is merely modified by this level.

The binary signal for each sampling period is available at the terminal 14 at some time during the sampling period when the difference output at terminal 8 exceeds the average difference level by at least the threshold amount. This binary output 14 may be stored in digital storage means, not shown, for a period of one picture element during which time a stored '1' digit may operate a visual or audible alarm circuit. It may also be fed to a display device showing the location of the object having at least the predetermined contrast excess over the average in relation to the remainder of the scene containing the object. For example, if a television camera, not shown, is used to view the scene and to produce the video output fed to input 5, the binary output at terminal 14 or a stored version of it may be superimposed on the video waveform of the camera before it is fed to a display monitor 16. If the scene is scanned by a detector, a synchronous image reconstruction device, not shown, may be provided to which the binary signal directly or a stored version of it may be fed, either alone or added to the detector output.

If the scene being viewed or scanned contains large areas in each of which the scene radiance changes slowly as a function of distance across the area, the difference signals at terminal 8 will be almost constantly small and hence the average difference signal level at terminal 12 will be low. The threshold level at terminal 15 will then be determined largely by the value of the reference level voltage at terminal 13 which may be set, for example, so that relatively small difference signals may exceed the threshold level if a high false detection rate is acceptable and it is desired to detect small contrast objects. Alternatively, the reference level voltage may be set so that only larger difference signals can exceed the threshold level. This alternative may be chosen if a low false detection rate is desired and if detection of only larger contrast objects is acceptable. However where scenes having a high content of sharp edged and/or highly contrasted objects are viewed, the average difference signal level at terminal 12 will be larger, on average, and the threshold level will be raised automatically.

Only objects contrasting enough to stand out against the relatively contrasting background will now be detected. The false detection rate will be reduced now over the rate which would have been obtained if the threshold level had remained at the value obtained with the scene of generally low contrast.

WHAT WE CLAIM IS: 1. Apparatus for processing a video signal comprising means for sampling the instantaneous value of the video signal at regular intervals of time, means for temporarily storing each sampled value, means for obtaining a differential signal as the difference between each instantaneous value of the video signal and the preceding stored sampled value, means for forming a running average of said differential signal related to a predetermined preceding time interval of said video signal, means for forming a running threshold level from the said running average of said differential signal, and means for providing an object indicator signal when the differential signal exceeds the running threshold level.

2. Apparatus as claimed in Claim 1 including a display device for utilising said object indicator signal to show the location of an object in relation to a scene giving rise to said video signal.

3. Apparatus as claimed in any preceding claim wherein said object indicator signal is of binary form, having one binary value when the differential signal exceeds said running threshold level and the other binary value when the differential signal equals or is less than said running threshold

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (6)

**WARNING** start of CLMS field may overlap end of DESC **. Such a time constant will give a decaying weight to earlier difference signals. Other forms of averaging circuit may be used in which difference signals earlier than a predetermined time are rejected altogether in forming the average difference signal. The output 12 of averaging circuit 3 is compounded by resistors R1 and R2 with a reference level voltage applied to terminal 13. The ratio of these resistors is chosen to provide a desired threshold level voltage at the junction of the resistors which is applied to the negative polarity input terminal 15 of comparator 4. Comparator 4 is a high gain differential input d.c. amplifier in which the output on conductor 14 is limited not to fall below zero volts and not to rise above a fixed positive voltage V. These two values 0 and V are taken in further processing and/or display circuits to represent the binary digits '0' and '1' respectively. If at any time therefore, the difference signal at terminal 8 is more positive than the threshold level voltage by a very small amount inversely proportional to the comparator gain, the output at terminal 14 will be driven to the upper limit V and a '1' digit will then be allocated to the corresponding picture element. If the difference signal is equal to or more negative than the threshold level voltage, the output at terminal 14 will be driven to the low limit zero and a '0' digit will be allocated to the corresponding picture element. The ratio of the resistors R1 and R2 and the reference level voltage applied to terminal 13 may be chosen so that threshold level voltage is dominated by the output level of the averaging circuit 3 or is merely modified by this level. The binary signal for each sampling period is available at the terminal 14 at some time during the sampling period when the difference output at terminal 8 exceeds the average difference level by at least the threshold amount. This binary output 14 may be stored in digital storage means, not shown, for a period of one picture element during which time a stored '1' digit may operate a visual or audible alarm circuit. It may also be fed to a display device showing the location of the object having at least the predetermined contrast excess over the average in relation to the remainder of the scene containing the object. For example, if a television camera, not shown, is used to view the scene and to produce the video output fed to input 5, the binary output at terminal 14 or a stored version of it may be superimposed on the video waveform of the camera before it is fed to a display monitor 16. If the scene is scanned by a detector, a synchronous image reconstruction device, not shown, may be provided to which the binary signal directly or a stored version of it may be fed, either alone or added to the detector output. If the scene being viewed or scanned contains large areas in each of which the scene radiance changes slowly as a function of distance across the area, the difference signals at terminal 8 will be almost constantly small and hence the average difference signal level at terminal 12 will be low. The threshold level at terminal 15 will then be determined largely by the value of the reference level voltage at terminal 13 which may be set, for example, so that relatively small difference signals may exceed the threshold level if a high false detection rate is acceptable and it is desired to detect small contrast objects. Alternatively, the reference level voltage may be set so that only larger difference signals can exceed the threshold level. This alternative may be chosen if a low false detection rate is desired and if detection of only larger contrast objects is acceptable. However where scenes having a high content of sharp edged and/or highly contrasted objects are viewed, the average difference signal level at terminal 12 will be larger, on average, and the threshold level will be raised automatically. Only objects contrasting enough to stand out against the relatively contrasting background will now be detected. The false detection rate will be reduced now over the rate which would have been obtained if the threshold level had remained at the value obtained with the scene of generally low contrast. WHAT WE CLAIM IS:

1. Apparatus for processing a video signal comprising means for sampling the instantaneous value of the video signal at regular intervals of time, means for temporarily storing each sampled value, means for obtaining a differential signal as the difference between each instantaneous value of the video signal and the preceding stored sampled value, means for forming a running average of said differential signal related to a predetermined preceding time interval of said video signal, means for forming a running threshold level from the said running average of said differential signal, and means for providing an object indicator signal when the differential signal exceeds the running threshold level.

2. Apparatus as claimed in Claim 1 including a display device for utilising said object indicator signal to show the location of an object in relation to a scene giving rise to said video signal.

3. Apparatus as claimed in any preceding claim wherein said object indicator signal is of binary form, having one binary value when the differential signal exceeds said running threshold level and the other binary value when the differential signal equals or is less than said running threshold

level.

4. Apparatus as claimed in any reced- ing claim wherein said video signal is de rived from a raster of line scans and said predetermined preceding time interval of said video signal is small compared to the duration of a line of said raster.

5. Apparatus for processing video sig nals substantially as described with reference to the accompanying drawing.

6. Apparatus for processing video sig- nals in combination with a display device substantially as described with reference to the accompanying drawing.