FR2687883A1 - IMAGE SIGNAL PROCESSING CIRCUIT. - Google Patents

Abstract

Provided is a circuit for processing an image signal which is produced by sampling an initial image at different sampling points along the frames and is output. It includes a decoder (1) which decodes the image signal on reception, a memory (2) which stores full image data of the image signal from the decoder, a switching circuit (4) which selectively extracts the full image data of the memory or the full image data of the decoder, and a control means (3) which serves to control the switching circuit.

Description

The present invention relates to an image signal processing circuit

  whereby, for example, when the transmission signal of a high definition television system is received on a receiver having the scanning parameters of the normal television system, the deterioration in the quality of the reproduced image is minimized. High definition television signal transmission systems having compatibility with the scanning parameters of the normal television system are well known. For example, the initial image can be sub-sampled using diagrams. sampling which each have different sampling points on each frame over several frames to form a period, and transmit the data thus subsampled In a receiver

  designed for the high definition television system, all sample points

  The initial data are reconstructed from the sampled data which is transmitted during a period so as to be displayed simultaneously, so that a high quality image is obtained. At the same time, with a receiver which has the scanning parameters of the normal television, the received transmission signal is displayed as it is, so that an image of normal quality is obtained. Techniques of this type are described in the Japanese patent.

  made available to the public under publication number 59-12 677.

  However, in the case where this type of television signal is reproduced by a receiver which has the scanning parameters of the normal television system, the fact that there is a difference for the sampling points at each frame causes a deterioration of quality of the reproduced image For example, in the well-known HD-MAC (High Definition-Multiplexed Analogue Component) system which is a high definition television signal transmission system having this compatibility, when the transmission signal of the system HD-MAC is received in a normal television receiver of the well known system D 2-MAC and is reproduced, a defect in the quality of the image is observed such as

  shaking of the edges of the image causing a disturbance at the edges.

  Furthermore, when the HD-MAC transmission signal of motion images with motion detection is reproduced on a television receiver of the D 2-MAC system, an unnatural movement is observed in the reproduced image, which is called the phenomenon of trepidation Furthermore, since, in the HD-MAC system, a transmission image is divided into blocks and the transmission signal undergoes a

  block processing, they can produce block distortions.

  To prevent these faults, several measures are taken before transmission in the HD-MAC coder. For example, the high frequency component of the image signal is attenuated by 6 d B in order to solve the problem mentioned first and the signal is submitted image processing to temporal filtering in order to scale the movement and thus solve the second problem mentioned. However, these faults cannot yet be completely eliminated and, when a receiver intended for the D 2-MAC system receives the transmission signal from the HD-MAC system, the image quality deteriorates which compares with the reception of transmission signal from normal MAC system

  in a receiver intended for the D 2-MAC system.

  The following description, intended to illustrate the invention, aims

  to give a better understanding of its characteristics and advantages; it is based on the appended drawings, among which: FIGS. 1 A to E are simplified diagrams showing sampling diagrams used in the HD-MAC system; FIG. 2 is a simplified diagram showing the method of transmitting pixels, or picture elements, in the HD-MAC system; FIGS. 3 A and 3 B are simplified diagrams explaining the defect resolved by the invention when a transmission signal from the HD-MAC system is reproduced by a receiver from the D 2-MAC system; FIG. 4A is a simplified diagram showing a first method of processing complete image data by an image signal processing circuit according to the invention; FIG. 4B is a simplified diagram showing a second method of processing complete image data by an image signal processing circuit according to the invention; Figures 5 A to 5 C are simplified diagrams explaining in detail the process of Figure 4 B; Figure 6A is a simplified block diagram showing a main part of the first embodiment of an image signal processing circuit according to the invention; and FIG. 6B is a simplified block diagram showing a main part of the second embodiment of a signal processing circuit

image according to the invention.

  To begin, we will explain, in relation to Figures 1 A to E, the sub-sampling carried out in the HD-MAC system In the HD-MAC system, a wide image, in 16: 9 format, is constructed at 1250 lines help

  and at a frame frequency of 50 Hz This image signal is under

  sampled so that the frequency band of the transmit signal is compressed Consequently, depending on the result of the motion detection performed, by block, on a block of 16 x 16 (lines) pixels, the diagram of sampling of each block suitably switches between three modes, namely an 80 ms mode (still image mode), a 40 ms mode (slow speed mode) and a 20 ms mode (large mode speed), so that the different sampling points of the initial image are sampled with each frame over a period of four frames, as shown in figures l A to l E Figures l A to 1 C show the diagrams of sampling associated with the luminance signal component, and Figures 1 D and E show the sampling patterns associated with the differential chrominance signal component, the numbers 1 to 4 which appear in the drawings designating the frame in which given them es of the respective points are sampled At the same time, with the D 2- MAC system, a normal image of 4: 3 format is constructed by means of 625 lines and according to a frame frequency of 50 Hz Consequently, to maintain compatibility with the D 2-MAC system, the subsampled data of the luminance signal component found in two lines close to the 80 ms mode and the 20 ms mode are interleaved and merged together into a single line, that the we

  emits This is called rearrangement of lines.

  Thus, taking the example of the 80 ms mode represented in FIG. 2, in the first frames of the first complete images and the first frames of the second complete images of each group of four frames, the pixel data being at the initial locations identified by O and the pixel data located at the initial locations marked by X, which are interleaved at the location marked by O between the locations marked by O at the time of

  the emission, are emitted In the second frames of the first images comprising

  full and the second frames of the second complete images of each group of four frames, the image data located at the initial locations marked with X, which are interleaved at the location marked with O during transmission, are transmitted in the 20 ms mode, the same sampling patterns of frame 1 of the complete image 1 and of frame 2 of the complete image 2, represented in FIG. 2, are used in the first complete images as well as in the second complete images This interlacing creates compatibility with the transmission constants of the D 2-MAC system having 625 lines In a television receiver of the HD-MAC system, the transmission signal is decoded so that the initial image is reconstructed , but, however, we will omit any

  provide a detailed explanation, as this does not directly relate to the invention.

  However, if there is an oblique rectilinear edge in an initial image, as represented by the real line connecting the points of coordinates 1-m and 13-a of FIG. 3 A, the pixels corresponding to the edge are actually arranged at the locations marked with * and M in the transmission system of the HD-MAC system When this signal is reproduced by a receiver intended for the D 2-MAC system, the line forming a straight edge is transformed into a zigzag, as shown in FIG. 3 B This is called the zigzag effect. In addition, in particular in the 80 ms mode, we can see by comparing the complete image 1 and the complete image 2 of FIG. 3 B that the pixels corresponding to the line of edge appear in different locations on the respective frames, which causes a deterioration of the quality of the image with a full frame rate of 25 Hz (period of 12.5 Hz).

board disturbance.

  To resolve these two image defects, complete image memories are used according to the invention, for example, to replace the pixel data. Since the edge disturbance appearing in a still image

  processed in 80 ms mode is caused by the difference in sample positions

  lonnage between neighboring complete images, if the data of neighboring complete images are identical, this edge disturbance does not appear Thus, according to this invention, the data of one of the two neighboring complete images are displayed in the other complete image also every two complete images, as shown in FIG. 4 A For example, the sampled data of frames n and n + 1 are displayed not only in the frames N and n + 1, but also in the frames n + 2 and n + 3, and the sampled data of the frames n + 4 and n + 5 are displayed not only in the frames n + 4 and n + 5, but also in the frames n + 6 and n + 7 Another method consists in displaying in the two complete images, as shown in the FIG. 4B, the averages of the sampled data found in the neighboring complete images For example, the average data of the frames N and n + 2 are displayed in the frames N and n + 2, and the data s

  averages of frames n + 1 and n + 3 are displayed in frames n + 1 and n + 3.

  With this latter method, in particular, not only the edge disturbance, but also the zigzag effect are reduced. Thus, as represented in FIG. SA, thus, as represented in FIG. 5A, the averages of the data of the pixels identified by E and Q in line 1, for example, of frame 1 of complete image 1 and of frame 1 of complete image 2, taking into account the fact that the pixel data identified by E in frame 1 of the complete image 1 and the frame 1 of the complete image 2 are initially the data of the pixels located at the locations marked by X located below, of two lines, practically present the values of the pixels located at the locations marked by *

in row 2.

  Likewise, the averages of the pixel data identified by E in line 3, for example, of frame 2 of the complete image 1 and of frame 2 of the complete image 2, taking into account the fact that the pixel data marked with E in frame 2 of complete image 1 and frame 2 of complete image 2 are initially the pixel data located at the locations marked by X above or below, practically have the values of the pixels located at the locations marked by E in line 3, namely at the locations actually emitted When there is a rectilinear oblique edge in an initial image, connecting the points of coordinates 1-m and 13-a as shown in FIG. 4, since the averages representing the pixels located at the locations marked by X are displayed at the locations marked by * in the frame 1 and the means representing the pixels located at the locations marked by * in the frame 2 are displayed at the locations indicated in FIG. B, the pixels corresponding to the edge are arranged at the locations marked with * In this way, the mark A shows the locations of the pixels having the means of the pixels corresponding to the edge and of the pixels not corresponding to the edge By observing FIG. 5 C, which shows a situation of overlapping of the weft 1 and of the weft 2, it can be seen that the zigzag effect is reduced due to the fact

  that we used the averages of the pixel data.

  To implement the first method mentioned above in an image signal processing circuit according to the invention, as shown in FIG. 6 A, the signal received from the HD-MAC system is supplied to a decoder 1 of the system Normal MAC (D 2-MAC), from which the luminance signal component Y and the differential components of the chrominance signal, ie U and V, are obtained The luminance signal component Y is supplied to a memory 2 of full image The luminance signal component Y from the MAC decoder 1 and the output signal from the full image memory 2 are

  supplied to motion detection circuit 3 as well as to a switching circuit

  tion 4 by means of which the luminance component signal Y coming from the MAC decoder 1 or the output signal from the memory 2 are selectively obtained in the form of a complete image in correspondence with a detection signal from the detection circuit 3 motion Thus, when a motion element is detected in the complete image, the output signal of the MAC decoder 1 is extracted as it appears, and, when a motion element is not detected, the image is considered as a still image and the output signal of the MAC decoder 1 is first extracted via the switching circuit 4 and identical data are simultaneously stored in the complete image memory 2 during the complete image period En second, the data is read from the full image memory and is extracted via switching circuit 4 during the next full image period, while the data need not be from MAC decoder 1 are stored in full image memory 2 during this period After that, the luminance signal component

  Y coming from the MAC decoder 1 and the output signal from memory 2 are alternating

  extracted at full frame rate The motion element can be extracted from motion information, in the form of blocks,

  which is included in the signal (DATV signal) of the HD-MAC system via the

MAC decoder 1.

  To implement the second method mentioned above in an image signal processing circuit according to the invention, as shown in FIG. 6B, the signal received from the HD-MAC system is supplied to a decoder 1 of the system Normal MAC (D 2-MAC), from which the luminance signal component Y and the differential chrominance signal components U and V are obtained The luminance signal component Y is supplied to a full image memory 2 as well only to a motion detection circuit 3 The luminance signal component Y coming from the MAC decoder 1 and the output signal from the complete image memory 2 are supplied to a complete image averaging circuit 5 Le luminance signal Y from the MAC decoder 1 and the output signal from the full image averaging circuit 5 are supplied to the switching circuit 4, through which the luminance signal component Y from the MA decoder C 1 or the output signal of the complete image averaging circuit 5 are selectively delivered, in the form of a complete image, in correspondence with the detection signal of the motion detection circuit 3 Thus, in this embodiment , the motion detection circuit 3 detects the motion element from the output signal from the MAC decoder 1, and, when a motion element is detected in the complete image, the output signal from the MAC decoder 1 is extracted as it appears, while when a motion element is not detected, the image is considered to be a still image and the output signal of the complete image averaging circuit 5 is extracted as it stands Motion detection can be obtained from motion information, which is in the form of blocks, via the MAC decoder 1, as in

the first case.

  In another variant, when the data contained in a complete image is replaced by the data of the previous complete image, it is possible, having regard to the first frame of each complete image, that the data of the pixels identified by El in the frame 1 of the complete image 1 and the frame 1 of the complete image 2 appearing in FIG. 2 are replaced by the data of the pixels identified by Q in the frame 1 of the complete image 1 and the frame 1 of the complete image 2 of the previous complete image In addition, when the data are replaced by their average, the average may not be the simple average of two pixels, but it may also be an interpolation datum obtained at from data of more than two suitable pixels, on the basis of a certain calculation formula. For example, we can use an average obtained from the data of the pixels located at the upper and lower lines of the

same frame.

  The D 2-MAC system and its decoder are well known, so we will omit to provide a detailed explanation The signal processing circuit

  image according to the invention can also be applied to a recording device

  While the case of the only luminance signal has been explained above, the invention is applicable to the differential chrominance signal. According to the invention, even if a high definition television signal from the HD-MAC system for example is received by a receiver having the normal scanning parameters of the D 2-MAC system for example, it is possible to obtain a quality of image which is identical to that obtained in the case where a normal television signal is received by a receiver having parameters of

  normal scans, like the D 2-MAC system.

  Of course, those skilled in the art will be able to imagine, from the

  circuit whose description has just been given by way of illustration only and

  in no way limiting, various variants and modifications do not depart from the scope of the invention.

Claims (5)

  1 image signal processing circuit intended to process an image signal which is produced by the sampling of an initial image at different sampling points according to the frames, and is transmitted, the circuit comprising: means decoding (1) for decoding the image signal at the time of reception; characterized by: storage means (2) for storing the complete image data of the image signal from the decoding means (1); switching means (4) for selectively extracting the complete image data from the storage means (2) or the complete image data from the decoding means (1); and control means (3) for controlling the means of switching (4).   2 Circuit according to claim 1, characterized in that said storage means (2) stores the data of a complete image coming from said decoder (1) every two complete images and said control means (3) controls said switching means (4) so as to extract, alternately at each frame, the signal from said decoder (1) or the data stored in said memory (2).   3 Circuit according to claim 2, characterized in that said control means is a motion detection means serving to detect motion information of the image contained in the image signal, and in that the switching means ( 4) is controlled by the movement information so as to extract the output signal from said decoder (1) when a movement is detected and to extract the signal from said decoder (1) or the data stored in said memory (2), each frame alternately, when a movement is not detected. 4 Circuit according to claim 1, characterized in that said means of   storage (2) is a complete image memory.   A circuit according to claim 1, characterized by means (5) for averaging a complete image, connected to said storage means (2) and to decoding means (1) and producing average data, and in that said means control (3) controls said switching means (4) so as to selectively extract the signal from said decoder (1) or the output signal   said complete image averaging means (5).   6 Circuit according to claim 5, characterized in that said control means (3) is a motion detection means for detecting motion information of the image contained in the image signal, and in that said means switch (4) is controlled by the motion information so as to extract the output signal from said decoder (1) when motion is detected and to extract the output signal from said averaging means (5)   full image when motion is not detected.   7 Circuit according to claim 1, characterized in that the image signal is transmitted in the form of a television signal from the HD-MAC system and in that the scanning parameters of the display means to be connected are those of system D 2-MAC.