SU1727210A1 - Method of suppression of pulse interference in video image signal - Google Patents

Abstract

The invention relates to the field of television technology and can be used to improve the quality 2 of the reproduced television image. The purpose of the invention is to reduce the time to suppress impulse noise. All the lines of the raster, except the first, are delayed for a time not exceeding the return time of the frame sweep, element by element is compared each element of each raw line with the corresponding element of the previous processed line, and if the comparison result is exceeded over a predetermined threshold, the current element of the raw line is replaced by the average value obtained from neighboring checked pixels. 1 il.

Description

(L

with

The invention relates to television technology and can be used to improve the quality of a reproduced television image.

The purpose of the invention is to reduce the time for the suppression of impulse noise.

The drawing shows a structural electrical circuit of a device implementing a method for suppressing impulse noise in a video signal of an image.

A device that implements a method for suppressing impulse noise in an image video signal comprises a multi-bit shift register 1, an adder 2, a memory register 3, a first switch 4, a first delay element 5, a key 6, a second delay element 7, a second switch 8, a third delay element 9, block 10 subtraction ,, source 11 reference voltage adder

12, calculator 13 module, comparator 14, pulse generator 15, frame pulse generator 16, first divider 17, third switch 18, second divider 19, second and third D-flip-flops 20, 21 and 22, respectively.

The method of suppressing impulsive noise in the video signal of an image is based on the fact that for the first decomposition line the average level is calculated over a certain neighborhood for each element of the first line, the average value is compared, the excess of the comparison result is determined by the specified threshold value, and if the comparison result is exceeded the value of the element to others closer to the average value. Then each previous line is delayed by the time of the horizontal scanning period in relation to

yu xj

Yu

about

for subsequent, successively comparing for each line each element of the image in each subsequent line with the corresponding element of the previous line, it is determined whether the comparison result exceeds the threshold value. If the comparison result exceeds the threshold value, replace the element of the subsequent line with the element of the previous line or with the average value adjacent to the replaced elements of the subsequent line. Due to the proposed method of eliminating impulse noise, for correcting the first line, it is necessary to perform (N + 1) -n, where n is the number of decomposition elements in the line, N is the number of points over which the signal of the decomposition element is averaged, adding hegerations: for correcting the rest, m- 1 lines, where m is the number of lines of decomposition, it is necessary to perform (N + 1) p addition operations, where p is the average number of pixels subjected to impulse noise. The total number of add operations will be

U (N + 1Hm-1 p.

The gain from the method will be (n + pm). The average value of p 10 ... 100 gain in speed will be 10-100 times.

A device that implements a method for suppressing impulsive noise in an image video signal operates as follows.

The pulse generator 15 continuously generates a pulse sequence with a frequency f 1. corresponding to the arrival frequency of the nominal picture elements. At the output of the frequency divider 17, a pulse sequence operates at a frequency f2 fi / r, where r is a number from a series of numbers 2, 3, 4, .... At the moment of the beginning of the image frame deployment, the generator 1.6 generates a frame sync pulse which zeroes the dividers 17 and 19 frequency and sets the single state of the trigger 20. The trigger 20 turns on the key 6, and the video signal from the device input is fed to the input of the multi-bit shift register 1. At the same time, the video signal is continuously fed to the input of the delay line 7. Since the trigger 21 is in the zero state, the control voltage of the switch 18 is absent and its output is switched from. the second entrance. Therefore, pulses with frequency fi are fed to the input of frequency divider 19 and synchronous inputs of register 1, adder

2, memory register 3. Under the influence of pulses, the shift register 1 fills its cells in the memory, and the value of the video signal of one nominal picture element falls into each cell. The frequency divider 19 divides the frequency of the input pulse sequence n times. Therefore, at the output of the divider 19, an impulse appears when all the cells of the shift register 1 are filled. This impulse promotes information in the chain of D-flip-flops 20, 21, 22, i.e. The second D-flip-flop 21 translates to a single state, and the flip-flop 20 is zeroed, since there is no information at its input. As a result, the key 6 disables the input of the device from the multi-bit shift register 1. The switch 18 is turned on, i.e. its output is connected to the first input, adder-normalizer 2

0 is also included, i.e. starts the operations of summing the input signals. At the same time, the frequency of advancing information in the shift register 1 and the frequency of filling the register 3 memory decreases from fi to f, in

As a result, the time interval between successive information advancement pulses is increased by a factor of r and the adder-normalizer 2 can perform more operations, for example, smoothing and weighing the components. The division factor g of the divider 17 is selected based on the known frequency f 1 of the element arrival, the images used, the element base used and the number of averaging elements of the image. The division ratio r cannot exceed the number of passive raster lines. When the image string is filled in

0 register 1, will be processed and the memory register 3 will be filled, and this will happen exactly after the n clock ticks of the signals in adder 2, the output of frequency divider 19 will have a pulse, which

5 will again shift the information in the D-flip-flop chains 20, 21 and 22: flip-flop 22 will be transferred to one state, flip-flop 21 will be reset. Switch 18 is turned off, and switch 4 is turned on. At this point, the output of the delay line 7 is the video signal, the first line of the image. This video signal is elementwise subtracted in subtraction block 1.0 from the smoothed video signal of the first line,

5 memory read from register 3 through switch 4 with a frequency d of the generator 15 pulses. At the same time, the smoothed video signal is fed to the first input of the adder 12, and the current value of the video signal of the line through the delay line 9 is fed to its second input. Differential bipolar

the signal from the output of subtraction unit 10 turns into unipolar in subtractor 13 module. The amplitude of this signal at each moment of time depends on whether the current value of the video signal of the image is above the average, i.e. averaged over the number of elements I video signal. Comparator 1-4 captures the moments when the modulus of the difference signal is above the average signal, classified as having a pulse noise, by generating a signal that includes switch 8. Therefore, the image element subjected to pulse noise is replaced by the contents of adder 12, which contains the average sum of two neighboring (already checked) image elements: previous in this line and having a similar location as in the previous line. The adjusted video signal from the output of the switch 8 is fed to the input line 5 delay. This video signal appears at the output of the delay line 5 just at the moment when frequency divider 19 counts n pulses, i.e. at a time equal to the horizontal period. Impulse. With the output of the frequency divider 19 promotes the contents of the chain of triggers 20, 21, 22; as a result, trigger 22 is zeroed and switch 4 is turned off. The video signal from the output of the delay line 5, cleared of impulse noise, through the switch 4 is fed to the second input of the subtraction unit 10, the first input of which starts to receive the video signal of the next image line after the corrected line. The circuit containing the subtraction unit 10, the calculator 13 of the module, the reference voltage source 11 and the comparator 14 perform element-by-element testing for impulse interference of the video elements of the image of the subsequent line using the previous corrected line as a reference. When a distorted pixel is detected, it is replaced by the current contents of adder 12. Thus, each subsequent line is compared with the corrected previous line for the rest of the raster. After the entire image has been scanned, the frame sync generator 16 extracts from the input mix of the video signal and sync pulses the instant of the start of the new image frame sweeping by generating a new frame sync pulse and the device operation cycle repeats.

The proposed method for suppressing impulse noise in an image allows to improve the quality of reproduction of the image on the screen of a video monitor by correcting impulse noise penetrating the communication channel between a transmitting television camera and a pi-5 control unit. The method makes it possible to reduce the required number of operations for image correction, and to carry out these operations in the time interval of the vertical sweep of the vertical scanning. As a result, an increase in the total (total) speed of the method of filtering impulse noise is expected - and the simplification of its technical implementation. The method is most appropriate to apply when developing 15 new image transmission systems of enhanced informativity, when the elimination of impulse noise using standard element base is not capable of providing the required speed, and the use of specialized element base sharply increases the cost of equipment.

Claims (1)

The method for suppressing impulse noise 5 in a video signal of an image, which consists in extracting a signal from each image signal from an image signal, extracting a signal from a video signal in a neighborhood of each element from the first line image. decomposition, calculate the average signal level of the neighborhood of each element of the first decomposition line, compare the average signal level of each element 5 of the image of the first decomposition line with the current signal level of each image element of the first decomposition line, determine if the comparison result exceeds the threshold value and 0, the threshold values are replaced by the current level of the image element signal with a level equal to the average value of the signal level of the neighborhood of the image element, ayu soup and that. With the aim of reducing the time of suppression of impulsive noise, simultaneously with the separation of the signal of each image element from the video signal, the signal of all image elements, except for the 0 image signal of the first decomposition line, is delayed by the frame time signal from the video signal pulse, the signal of each image element of the previous element is compared by element , starting from the first, the line of decomposition with the signal of the adjacent element of the image of the current line of decomposition, and when the result of comparison is exceeded the threshold value They replace the value of the current pixel signal with the signal value received. by averaging averaging the sigma value of the signals of the preceding element of the image element in the image items - in the current line of the previous decomposition line, and decomposition.