BACKGROUND OF THE INVENTION

This invention relates to motion detection in a television picture that is contained in a television signal and more particularly to a motion detector which employs information included in the frequency spectrum of the television signal.

The prior art is . aware of the fact that a television signal contains information which is relative to the motion in the television picture content. Such techniques have been used in intrusion detection or video surveillance systems where it is required to be able to detect motion by examination of a video signal. A known method employed in the prior art compares a real time image with a delayed image from a field or frame store. In any event, such systems require a relatively expensive frame store or memory which stores the video signal of an entire picture frame or-field. The video signal that contains the • picture information of a subsequent picture frame is compared with the stored video signal and the result of such comparison provides a signal indicative of motion.

Essentially, such systems operate on the fact that motion causes the content of a pixel in a present video signal to be different from the corresponding pixel in the stored frame. In any event, as one can imagine, these systems have difficulties in that an expensive frame store is required. A further difficulty with such systems is that they are affected by changes in illumination which can cause false alarms.

It is also known that motion information is contained in the frequency spectrum of a television signal. See, for example, U.S. Patent No. 4,641,186 issued on February 3, 1987 to D. H. Pritchard and entitled "Motion Detector That Extracts Motion Information From Sidebands Of A Baseband Television Signal" and assigned to the RCA Corporation.

-2-

As described in that patent, there is a motion detector for detecting motion in a picture contained in a television signal which includes a filter that extracts a portion of the frequency spectrum of the television signal. The amplitude of the side bands of the television signal in this portion of the frequency spectrum is indicative of motion in the picture scene. The signal from the filter is coupled to a peak detector that produces a motion indicative signal. The amplitude of the motion signal is proportional to the absolute value of the amplitude of the sidebands in this portion of the spectrum.

As one can ascertain from the above patent, the motion detector as described is particularly useful when such television apparatus includes a line store memory for adapting the interlaced signal for display in a non¬ interlaced scanning format. As is known, the adaptation to a non-interlaced scanning format is provided to reduce certain conditions that are caused by displaying a picture using the interlaced scanning format at the television receiver. It should be apparent to those skilled in the art that there are many additional applications _τ apart from surveillance, where it is required to be able to detect motion by examination of a video signal. There are also systems used which employ digital video signals. These signals are processed by computer techniques to achieve picture recognition. In any event, the systems require extensive compression techniques to accommodate the high data rates of a television camera and utilize complicated computer programs to process data. In view of the above, it is indicated that the apparatus employed with such detection schemes should be economical, easy to implement and be reliable in operation. In this manner, the television signal can be adequately employed to detect motion and, hence, act in conjunction with an intrusion detection system or other system to indicate the presence of motion.

- -

SUMMARY OF THE INVENTION One aspect of the present invention is to provide a motion detector that employs information contained in the frequency spectrum of a television signal to produce an output indicative of motion.

It is a further aspect of the present invention to provide a motion detector which employs a narrow bandpass filter at an harmonic of the vertical sweep frequency. The output of the filter results in a carrier frequency signal which is amplitude modulated during the presence of motion.

It is a further aspect of the present invention to provide an inexpensive and reliable motion detector by extracting motion information indicative of sideband components at the vertical scanning rate and utilizing typical AM demodulation techniques to detect such motion.

BRIEF DESCRIPTION OF PREFERRED EMBODIMENT A motion detector apparatus of the type employing information contained in the frequency spectrum of a television signal used for providing a television picture which frequency spectrum for a stationary television picture consists of a series of components at harmonics of the vertical and horizontal scanning frequencies and wherein during motion in said television picture said series of components contain additional sidebands comprising narrow band pass filtering means responsive to one of the components of said vertical scanning frequency to provide at an output a sinusoidal signal at a given multiple integer of said vertical scanning frequency which signal will be amplitude modulated only during motion in said television picture and means coupled to said band pass filtering means for detecting said amplitude modulation to provide an output signal indicative of said motion.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 illustrates a part of the spectrum of a television signal which covers the range up to 20 kHz;

FIG. 2A illustrates a spectrum of the fundamental component of the field scanning rate and two associated harmonics;

FIG. 2B illustrates the field scanning rate spectrum, as affected by motion;

FIG. 3A illustrates the field scanning rate for a stationary picture while FIG. 3B shows the field scanning rate for a picture having motion depicting 1 Hz sideband components due to such motion;

FIG. 4 shows the frequency response of a filter utilized in conjunction with this invention; FIG. 5 shows a simple block diagram of a motion detector according to this invention;

FIG. 6 shows a detailed block diagram of a motion detector employed in this invention;

FIG. 7 shows a series of timing diagrams depicting the outputs at various points designated in conjunction with FIG. 6;

FIG. 8 shows pertinent timing waveforms depicting the outputs of the various circuit modules of in FIG. 6 in regard to motion and no motion conditions; and FIG. 9 is an actual circuit schematic diagram of a motion detector, as for example depicted in FIG. 6.

DETAILED DESCRIPTION Referring to FIG. 1 there is shown a schematic diagram illustrating a part of the spectrum of a television signal which covers the range up to 20 kHz. It is well known that the spectrum of a television signal for a stationary picture consists of a series of components at harmonics of the horizontal and vertical scanning frequencies. In FIG. 1 the horizontal scanning frequency is represented by the term f^ whereas the vertical scanning frequency is represented by the term f _v . In any event, the components of major magnitude

- - are the DC component, the field frequency components and the components of the line frequency and its harmonics.

As seen in FIG. 1, surrounding each line frequency harmonic is a cluster of components each separated from the 5 ^' J next by an interval equal to the field scanning frequency. In FIG. 1 the field scanning frequency or vertical scanning frequency is designated as 60 Hz for the NTSC system and (50) Hz for the PAL system.

Referring to FIG. 2A there is shown the fundamental

10 component of the field scanning rate f _v and two of its harmonics designated as 2Xf _y and 3Xf _v . As can be seen from

FIG.. 2A, one again is concerned with the NTSC system wherein the field scanning rate f _v is at 60 Hz with the second harmonic at 120 Hz and the third harmonic at 180 Hz. As one

15 can ascertain from FIG. 2A, the amplitudes of the harmonics have constant values depending on the scene content.

However, when motion is present the spectral components contain additional sidebands due to the change of information from one field to the next. Essentially, the field spectral

20 component is amplitude modulated by motion in the televised scene as shown, for example, in FIG. 2B.

For purposes of further explanation, let us assume that one sees a dark grey raster having a narrow horizontal white bar that is in constant motion between the top and 25. bottom of the television picture. ^"" If the bar is moving at a speed of two picture heights per second then the bar will only have moved l/30th of a picture height (525/60 standard) , between successive vertical scans and will return to its starting position and direction of motion after one second. 30 By referring to FIG. 3A and 3B, one can determine that the spectrum of the resulting video signal will be characterized by having 1 Hz sidebands due to this motion. FIG. 3A essentially shows the fundamental component of the vertical sweep frequency without motion, while FIG. 3B shows 5 the fundamental component of the vertical sweep frequency with motion and with the 1 Hz sideband components that are due to the motion. In general, the amplitude and spread of the motion sidebands will depend on the relative size and

-6- speed of the moving area. However, the above example shows that motion of practical significance always produces very low frequency sidebands on either side of the field scanning rate. 5 ^" Referring to FIG. 4 there is shown a typical band pass characteristic of a filter which can be utilized to detect motion by responding to an harmonic of the field scanning rate or the vertical sweep frequency. As seen in FIG. 4 the filter basically has a Q equal to 40. Essentially

10- the filter is offset from the second harmonic of the field scanning rate by 1.5 Hz as a typical example. The filter has a 3 db> bandwidth which is approximately equal to 3 Hz and the characteristics of the filter are shown clearly in FIG. 4. It is, of course, understood that one can implement the

15 filter characteristics shown in FIG. 4 by many conventional techniques, including the use of operational amplifiers employing RC selective filter circuits and many other filter techniques. It is, therefore, indicated that one skilled in the art' could easily implement a filter having the

20 characteristics depicted in FIG. 4.

It is clear that motion can be detected by using a narrow bandpass filter having the bandpass characteristics shown in FIG. 4. As indicated the bandwidth of the filter is about 3 Hz but bandwidths between 3-5 Hz will operate as

255 well. The filter basically has a center frequency located near the second harmonic (120 Hz) of the vertical scanning rate. The second harmonic is utilized due to the fact that there would be less disturbance from power supply hum at this frequency than would occur if the actual scanning frequency

30. of 60 Hz were employed.

In any event, the bandpass center frequency is selected at an integral multiple, n, of the vertical scanning frequency.

As one can ascertain from FIG. 4, the center

35 frequency of the filter is offset by approximately 1-2 Hz from the vertical scanning harmonic to emphasize the motion

- - components with respect to the main fixed signal component. Ass seen from FIG. 4, the particular offset selected is r.5 Hz-

Referring to FIG. 5, there is shown a simple block 5 ^" diagram of a system which employs the above-noted effects to detect motion. Essentially, the video camera 10 supplies a television signal which is applied to the input of the bandpass filter 11 having the frequency response depicted in FIG. 4. The television camera 19 is set up to monitor, an area 10. under surveillance and hence provides a television signal indicative of the picture content of the area being monitored.. The output of the bandpass filter 11 will be a 120 Hz sine wave having a typical amplitude of about 40 V peak to peak for a typical scene being monitored by the

15 camera 10. The output of the bandpass filter 11 is coupled to the input of an AM demodulator 1 . The AM demodulator is a typical amplitude demodulator, many examples of which exist in the prior art. The effect of motion, as indicated above, results in amplitude modulation of t e 120 Hz sine wave which 0 serves as a carrier frequency. This modulation is detected by the AM demodulator 12. The output of the AM demodulator 12 is applied to the input of an interface logic circuit 14 which will be more thoroughly described. As seen, the output of the interface logic circuit 14 will provide a high signal 5; or a first output for the presence of motion and will provide a low signal or a second output indicative of no motion. It is, of course, understood that care must be taken in regard to the circuit implementation due to the extremely small amplitudes of motion information that exists relative to the 0 carrier frequency.

Referring to FIG. 6 there is shown a block diagram of a motion detector which operates according to the above- described principles. It is noted at the outset that FIG. 6 will be explained in conjunction with the wave forms depicted 5 in FIG. 7.

FIG..7 depicts the various wave forms indicated by the encircled numerals as, for example, 1 through 9 shown on the block diagram of FIG. 6. In this manner, FIG. 7 depicts

the various waveforms and their properties as developed by the circuitry depicted in FIG. 6.

The video input signal from a camera or other device is applied to the input of bandpass filter 20. It is 5.. again indicated that filter 20 has the frequency response as shown in FIG. 4. The filter 20 will provide, at an output, the sinusoidal signal which essentially is at a frequency of 120 Hz. As indicated above, the amplitude of the sinusoidal signal emanating from filter 20 and depicted in FIG. 7-1, has 0 a constant value which magnitude depends on the scene content. If motion occurs, the sine wave at the output of filter 20 will be amplitude modulated. Motion can then be detected by responding to the envelope of this modulated signal. Thus, as shown in FIG. 6, the output of the 5 filter 20 is directed to an input of a full wave rectifier or bridge rectifier 21. The output of the bridge rectifier 21 is shown in FIG. 7-2. The output of the full wave rectifier 21 is directed respectively to the inputs of a first and a second low pass filter as 22 and 23. The low pass filter 22 D? has a shorter time constant than the low pass filter 23. In this manner low pass filter 22 is able to follow the modulated envelope. The output of the low pass filter 22 is shown in FIG. 7-3. As one can ascertain from FIG. 7-3, the output of the low pass filter 22 contains some 240 Hz ripple 5 due to the shorter time constant. In any event, the output of low pass filter 23 is shown in FIG. 7-4. As one can ascertain, due to the fact that the time constant of low pass filter 23 is longer, it suppresses the modulation but essentially follows the DC value of the rectified signal as ø; shown in FIG. 7-4.

The signals at the output of each low pass filter are coupled to respective inputs of a comparator circuit 24. The comparator 24 produces a signal at its output which is a 240 Hz square wave signal as shown in FIG. 7-5. With no 5 motion, as seen from Fig. 7-5, the square wave output from comparator 24 has a duty cycle of about 50%. If motion occurs, the duty cycle changes or the signal at the output of comparator 24, becomes a pulse width modulated signal.

There is an extreme advantage in using the two filter method. Namely, by employing low pass filters 22 and 23 the following operation occurs. The low pass filter 23 is used to provide a DC reference level for the signal emanating from filter 22. Essentially the signal output of low pass filter 23, having the longer time constant, is used as the slicing level for the signal output from low pass filter 22. The reason that this technique is preferable is in this manner the duty cycle of the comparator output signal is independent of the amplitude of the sine wave carrier signal which signal is the resultant signal obtained from a stationery picture. Because the slicing level is automatically adapted by filter 23 having the longer time constant, one then provides a comparator output signal which is relatively independent of the sine wave amplitude for a stationary picture. The use of the dual low pass filters as 22 and 23 will also compensate for any ^' changes due to temperature drift. In any event, the duty cycle of the comparator output signal is extremely sensitive- to any change in amplitude of the sine wave carrier signal which results from motion. This occurs because the slicing level or reference level emanating from the low pass filter 23 is too slow to track the change in carrier amplitude due to motion. Hence, the output of the comparator 24 essentially can be used as an indication of motion based on the change of the duty cycle of its output waveform. In any event, one still requires the system to provide a go, no go output. Hence, the system must provide an output which either says there is motion or there is not motion. Thus, the output of the comparator is applied to a first input of an exclusive NOR gate 26. One input to the exclusive NOR gate 26 is applied directly from the output of the comparator 24 while the other input from the comparator 24 is applied through a delay circuit 25. The output of the delay circuit is shown in FIG. 7-6 while the output of the exclusive NOR gate 26 is shown in FIG. 7-7.

-10-

The delay circuit 23 provides a delay (T ₃ ) and is implemented by utilizing a series connection of logic gates. The delay ₃ is typically longer than 40 nanoseconds (ns) due to the fact that the circuitry, as will be shown, employs TTL logic. This logic has speed limitations. in any event, the maximum value of the delay is- also limited by the 240 Hz signal used in this circuit. In practice, a value of 500 ns is reasonable. As indicated, the output signal from the exclusive OR gate 26 is shown in Fig. 7-7. Immediately CT beneath Fig. 7-7 there is shown an expanded diagrammatic representation of this signal to indicate the triggering edges for first and second one shots designated by reference numerals 27 and 28 respectively.

The output signal from the exclusive NOR gate 26

15 results from detection of the transients of the square wave signal output from the comparator 24. This involves detecting when the signal goes low to high' or high to low. As indicated, the wave form of FIG. 7-7 has below it an expanded diagram showing the edges which trigger the

20 respective one shots 27 and 28. Thus, as seen, the signal from the output of the exclusive NOR gate 26 operates to trigger the one shot 27 via a transition from negative to positive. The output of one shot 27 is shown in FIG. 7-8. The output of the one shot 27 is applied to the B input of

25 the one shot 28 while the A input of one shot 28 receives an input signal from the output of the exclusive NOR gate 26. The one shot 28 is connected so that it can only be triggered by the negative going edge (the transition from positive to negative) of the signal applied to its input A. If the

3 signal at the input B is high then one shot 28 can be triggered. In any event, if the input B is low then one shot

28 cannot be triggered and it will have an output which will always be in a low state, as for example shown in FIG. 7-9.

^• This state is indicative of no motion. It is noted that the

35 time constant of one shot 27 is chosen to have a value which is just shorter than (2.084 ms - T ₃ ) . This value can be adjusted for by means of potentiometer 9. In this case, one shot 28 cannot be triggered because the falling edge occurs

always at a time when the other input is low. Therefore, the output is in a stable low state with no motion, as indicated by wave form 7-9.

Referring to FIG. 8 there is shown timing diagrams designated again by encircled reference numerals 5, 6, 7, 8 and 9 which correspond to the encircled reference numerals of FIG. 6 and FIG. 7. The diagram of FIG. 8 shows on the left hand side the various outputs as described above (FIG. 7) for no motion and at the right hand side the various diagrams indicative of motion. Hence, as one can see from FIG. 8-5, the output of comparator 24, during the presence of motion, changes from a square wave having a 50% duty cycle to a pulse width modulated wave. In this manner the delayed signal at the output of delay 25 is also shown in FIG. 8-6 and essentially corresponds to the signal at the output of the comparator as delayed by the delay T3.

The waveform shown in FIG. 8-7 depicts the output of the exclusive. NOR gate 26 for no motion (left) and then for motion (right) . As one can ascertain from FIG. 8-7 the number of pulses at the output increases for the presence of motion as the duty cycle changes. The wave form shown in FIG. 8-8 depicts the output of the one shot 27 both for no motion and motion and the waveform of FIG. 9 shows how the output of one shot 28 transforms from a low value to a high value indicative of motion detection. Essentially, as one can determine from FIG. 8 and from the above-noted description of FIGS. 6 and 7, what occurs during the presence of motion is as follows. Due to the mismatch in the slicing level applied to the comparator for motion, the output signal from the comparator is pulse width modulated (8-5 and 8-6) . When the pulse width changes the positive going edges approach each other reaching a point where they begin to trigger the one shot 27. Once this occurs, the one shot 28 can also be triggered because the signal at its input B is now high when the negative going edge at the input A arrives. The time constant of the second one shot must be longer than 2.084 miliseconds so that it can be retriggered. As long as

- motion takes place, the output of the one shot 28 remains in the high condition. This output, being high, can then be utilized to activate an alarm as indicated in FIG. 6.

Referring to FIG. 9, there is shown a detailed circuit schematic of the motion detector depicted in the block diagram form of FIG. 6. Each of the representative modules, such as the filter 20, the rectifier 21 and other associated modules of FIG. 6, have been designated by the same reference numerals. In any event, the circuit shown in FIG. 9 incorporates the actual components as well as component values necessary to implement the circuit operation and structure depicted in FIG. 6. Essentially, as one can ascertain from FIG.9, the active filter portion consists of two similar stages, each employing an operational amplifier designated by IC-1 and further referenced by reference numerals 50 and 51. The configuration of the operational amplifiers 50 and 51 operating in a band pass configuration is well known. The frequency response of the filter, consisting of the two operational amplifiers 50 and 51, and their associated components can be adjusted by adjusting the potentiometers as 52 and 53. It is further noted that in FIG. 9 all component values have been designated by suitable component notations on the figure.

The rectifier circuit, which is circuit 21 of FIG. 6, contains two transistors 60 and 61, each being a BC 107, having their emitters and collector electrodes connected together and receiving inputs to the base electrodes. The filter output signal emanating from the output of operational amplifier 51 is applied to the base of transistor 60 directly and is applied to the base of transistor 61 via an operational amplifier circuit 52. The circuit 52 is of a unity gain configuration and, therefore, operates as an inverter. Hence, during the positive half of the sine wave one transistor conducts while the other one remains in cutoff. During the negative half of the sine wave the conducting transistor goes into cutoff and the other transistor conducts. This sequence of operation continues. In this manner, a rectified signal appears at the common

emitter output of both transistors, which is further indicated by the encircled reference numeral 2 which corresponds to FIG. 7-2.

A CMOS exclusive OR gate arrangement is employed to 5 _. produce the delay as delay 25 of FIG. 6. The output of the first gate in the delay circuit is loaded by a 10 NF capacitor to increase the gate delay. As indicated above, the time constant associated with the delay must be long enough to be accepted by the one shot 27. For each of the 0" one shots a TTL integrated circuit is selected because of its improved temperature stability. In any event, as indicated in~ FIG'. _* 9, the IC's utilized are as follows: IC-1 is an RCA CA 084 integrated circuit; IC-2 is an RCA CA 311 integrated circuit; IC-3 is an RCA CD 4077 integrated circuit; and IC-4 is a Texas Instrument 74LS/23 integrated circuit. Thus, as one can ascertain from the above description, the circuit employed utilizes conventional integrated circuit components which are available as off-the-shelf components. In any event, 'the circuit operates to detect motion by means of detecting the amplitude -of sideband frequencies at a harmonic of the vertical sweep frequency.

As indicated, the above-noted description involved the use of the second harmonic of the vertical sweep frequency, namely 120 Hz. In any event, it should be apparent to those skilled in the art that one could utilize the third harmonic at 180 Hz or any suitable harmonic. It is also understood that if one guarded against power supply hum, or power supply interference, one can employ the techniques described above, directly at the vertical sweep frequency, namely 60 cycles for the NTSC system or 50 Hz for the PAL system.