FR2513056A1 - Video processing circuit for telephone line transmission - uses microprocessor and comparator to determine difference from previous image for transmission of difference signals - Google Patents

Abstract

The video signal is applied to a fast A/D converter (1) whose output is sampled in time periods for application to a dynamic RAM (4). The cycle time for the RAM is less than the sampling period. A line and frame sync. separating circuit (3) operates in conjunction with an interface (9) for the memory to ensure that the memory is refreshed and that the sampling clock is controlled. A microprocessor based processing circuit (5) with peripheral units reads the memory registers corresponding to the different areas of an image and determines which have varies more than a predetermined level compared with a previous image. This process is aided by a comparator (7) and an authorisation system (8) which selects and accumulates data for the areas for comparison. The differences only are transmitted through a matching interface for the telephone line.

Description

Video image processing method allowing reduction
variable time information and transmission.

 The present invention relates to video image processing, and it relates more particularly, but not exclusively, to an image processing method for its transmission in variable time on a communication medium.

 In a video signal transmission system on a communication channel, it is often necessary to reduce the number of images transmitted in order to limit the width of the band necessary for the transmission, to that of the communication channel. The most common application for transmitting video signals with bandwidth reduction is the transmission of video images over a self-switched telephone line.

 The width of the passband of such a communication channel being of the order of 3 kilohertz, it is necessary to perform frequency compression in a ratio close to 2,000.

 It is therefore no longer possible to transmit 25 images per second defined on 625 lines, as in normal video, but 12 complete images per minute defined by I28 points per line on I28 lines, or 3 complete images per minute defined by 256 points per line on 256 lines, with point-to-point transmission. The images are memorized at well-defined instants, and unrelated to the variations thereof.

 Correlation techniques, coding methods make it possible not to transmit all the points of an image, but the successive differences between neighboring samples; or an interframe coding by which only those parts of the image which change from one frame to the next frame are transmitted (evening introduction to patent n-2.28I.023), and thus reduce the information rate of a transmitted image, therefore also its transmission time.

Another application is that of storing video image sequences. In the above examples,
We have the possibility of recording on an audio medium with a bandwidth of 3 kilohertz (for example: a magnetic tape), the output of the modulator of the transmitter. Reading this medium at the input of the receiver's demodulator will make it possible to replay the sequence of compressed video images at 3 kilohertz. The patents n'2.315.813 and n-2.386.2I3 report problems of transmission time of video signals on a communication channel.

The present process which is the subject of the invention makes it possible to reduce the information of a sequence of video images as well as the capacity of the systems for memorizing this sequence. I1 also makes it possible to reduce the time of transmission -sion of video signals on a communication medium with reduced bandwidth compared to that of the signal
normal video.

 The video image is structured in elementary zendes, each zone representing any set of points belonging to the image.

 The choice of the structure of these zones is not limiting and cannot therefore limit the industrial property of the invention.

 We can structure for example the video image either in vertical or horizontal bars, or in rectangles of different dimensions (plate 1/3 fig.1).

 The processing consists in memorizing the initial video image of the sequence considered, then of comparing it with the following images of this sequence. The comparison of the two images is made between corresponding zones which must keep the same structure.

 In the case where the comparison does not reveal any variation in the area, the image initially memorized remains the same, and the comparison continues with the following images. Otherwise, the modified zones replace the corresponding ones in the memory and thus, the image is updated in these regions considered. The comparison continues between the updated memorized image and the following images.

 The capacity of the information of an image of the video sequence can vary between zero, in the case of invariance of the image compared to the previous memorized, and the total capacity of the image, neither the latter varied on all areas.

 Consider for example a sequence of four images such that: the second is identical to the first; the third differs from the second by half; the fourth differs from the third by a quarter.

 For normal storage of these four images, each will occupy a quarter of the total memory capacity. The second image will occupy a quarter of redundant information in memory; the third image will occupy I / Se; and the fourth 3 / I6e.

 There will be for this sequence 9 / I6e of the total capacity of the memory used in a redundant way.

 According to the process which is the subject of the invention; for a total memory capacity identical to that mentioned above; allocation will be: 1/4 of the capacity for the first image; no capacity for the second image; I / Se of the capacity for the third image and 1 / I6e of the capacity for the fourth image. The sequence of these four stored images represents 7 / fie of the total capacity used previously.

 The reduction of the information of a sequence of images according to the method which is the subject of the invention also makes it possible to reduce the time of transmission of video signals on a communication medium.

 For a defined capacity of a communication medium, the transmission time of a set of data is proportional to the number of these data. The reduction in the number of data results in a reduction in the time for transmission.

 The method which is the subject of the invention makes it possible to store the image representative of the subject or of the object which has caused an image modification of the sequence.

 Example of a device which does not limit the industrial property of the invention: The device for implementing the method which is the subject of the invention is a system for transmitting video images on self-switched telephone lines carrying out the comparison processing - its of the image to be transmitted at the level of the transmitter, and the updating of the image, of which the modified zones have only been transmitted, at the level of the receiver.

 Such a system is composed: (plate 1/3 fig.2).

 - A transmitter <E) performing the frequency compression of a video signal (for example: from a camera 1), the processing of this signal according to the process object of the invention, the modulation of the compressed video signal on the phone line.

 - A receiver (R) ensuring the demodulation of the transmitted signal, the restructuring of the image and the conversion thereof, into a standard video signal for reproduction, for example on a video monitor (2).

Figure n'3 on board 2/3 shows the block diagram of the transmitting part of the system. It consists:
- In 1: A fast analog-digital converter
(I5 MHz.) 5 bits.

-) In 2: A time sampling of the output of the
converter 1. (Sampling period of
200 nanosecond word).

- In 3: A line and frame synchro extractor
video signal for synchronization
moire.

- In 4: A dynamic MOSRAM memory with
cycle time less than 200 nanoseconds.

- In 5: A processing unit based on a micro-
assignee and its peripheral elements.

- In 6: An interface for adapting the autocom line
mutated.

- In 7: The following comparison system
a mode of the process which is the subject of the invention.

- In 8: A comparison authorization system
including: - A sensitivity comparison comparison.

- A selection of areas to raise awareness
for comparison.

- Accumulators for information
comparisons.

- In 9: An interface for memory management and
its refresh, and the sampling clock.

Channels: (a) Represents video signal output
stored numeric values.

(b) Represents video signal output
digital samples.

(c) Represents the memory command
of the sampled image.

The transmitter has the rdle: (plate 2/3 fig.3).

- The realization of the analog-digital conversion
than at high speed (1) on 5 bits of the video signal
present at its entrance; each 5-bit word represents
feeling an eraduation of gray among 32 levels.

- To carry out a sampling in time (2) of the information
previous digital training, with a period of
the order of 200 nanoseconds to determine 256
points per video line.

- The possibility of memorizing for the duration of a
weaves a 256 line image into a fast memory
of dynamic RAM type having nn cycle time
less than the previous sampling period
This memorization is done on order of the unit of
treatment (5). For a correct recording of
the image in the memory, the addresses of this last
are generated from the sample clock
lonnage (included in 9) and extracted synchros
of the video signal (3).

- Comparison of 5-bit words from the sample
scaling of the current digital video image with
those of memory reading synchronized with
sampling (7). The state of a logic signal
at the output of the comparator defined equality or indi-
equality of the two words compared. A system (8) of
validation of the comparison and memorization of
comparator output states used to define
the structure of the image areas, possibly
allow comparison only in certain areas
sensitive to the image, the accumulation of data
of the comparator output in different registers
each corresponding to a defined area, and for
see raise awareness of variation by taking into account
from all points in the area, or that from a certain
significant number of these points.

- The determination of the zones having varied at the end of
the comparison between the two images. This is real
read in the processing unit (5) which reads the
registers corresponding to the different areas of
the image, and decides that this area has exceeded a threshold
variation, this area is updated in the
memory upon arrival of the next image on transmission
tor and the procedure for transmitting this zone
begins with issuing the location address
lie in the memory of this area, then by trans
mission series of these points constituting it from
slow reading of the relevant part of the memory
and converting this data into signals suitable for
size of the communication line.

The memory is in almost permanent reading interspersed
brief recordings for updating
the image after a zone variation detection.

- Adapt the transmission of data to the communication line
munication (6).

Figure n04 of plate 3/3 represents the block diagram of the receiving part of the system. I1 consists of: - In 1: A telephone interface ensuring adaptation
and maintaining the line.

- In 2: A processing unit based on a micropro
assignee and its peripheral elements.

In 3: A dynamic MOSRAM memory, with
cycle time less than 200 nanoseconds.

- In 4: A video sync generator controlled by
quartz.

 - In 5: A memory management and refresh interface.

 - In 6: A fast digital-video converter.

The role of the receiver is: (plate 3/3 fig.4).

- Reception of data by a line interface (l),
- Decoding in the processing unit (2 of the addresses
itself from the location of the written area, and so the
selection bn recording of the relevant part of
memory (3). (management of the latter by 5).

- Recording in memory of the data from
of the line, converted in the processing unit
(2) synchronously with the arrival of the latter
res.

-4The operation of the memory in fast reading
almost permanent, and the conversion of the digital output
risk of this in analog signal to which is
added synchronization (4) for the realization of
standard video signal. (The conversion takes place in
6) Processing unit of the transmitter and receiver
is programmable because it is carried out around a micropro
stopper, and allows multiple versions of function
detection and transmission.

System features:
- The image is defined by 256 points per line on 256
lines, on 32 gray levels (signal coded on 5
bits).

 - Transmission of the complete image in 19 seconds.

-Image structured in I6 vertical areas of the same
dimension.

 - Transmission of an I-zone, 2 seconds.

- Transmission of an image of a variable sequence
between 1.2 and 19 seconds.

 At the start of service, the transmitter transmits a complete image to the receiver, then performs the zone-to-zone comparison of the image arriving with that memorized.

If it does not find a variation, it remains in permanent comparison on the successive images with that recorded and emits nothing on the line. The receiver having received an image and nothing else after, keeps it in its memory.

 As soon as the transmitter has detected a variation in an image zone, it immediately updates its memory and transmits the variation to the receiver, which proceeds to replace the zone in question. At this time, transmitter and receiver have the same stored image, and the comparison cycle starts again at the transmitter.

 The method which is the subject of the invention can be applied everywhere or a reduction in the information of a sequence of images to be stored or transmitted is necessary.

 I1 can allow an increase in the number of multiplexed video channels on one communication medium.

 It can also constitute a means of storing more information, for archives, or data banks, or for memory media of large diffusions such as videocassettes and videodisks with the aim of saving memory capacity.

 The method which is the subject of the invention also applies to the reduction and transmission of information belonging to various technologies which allow the storage and comparison of the set of information constituting sequences.

Example not limiting the industrial property of the invention:
- You can memorize a set of words without> aemo-
laugh at the repetitive words in this set; the pro
assigned can therefore be applied to a synthesis system
or speech recognition.

- A particularly interesting video application
of the process which is the subject of the invention, is remote monitoring
launch remotely via telephone lines.

- Reception of more images in a time
given to follow an action, allows a better
understanding of it, and in the case of a
monitored room in which there should not be
of movement (for example: a vault during
at night), the slightest element entering the
camera field will be detected and stored
in the transmitter.

 The transmitter therefore stores the image to be transmitted in synchronization with the movement that has triggered the comparison, and can then command a line seizure and a transmission of this image to the corresponding receiver.

Claims (9)

 1. Process for processing video sequence images, in particular for recording and / or transmitting a sequence of video images on a medium of limited capacity, of the type comprising the structuring and storage operations of characterized images in that it consists in successively combining the following operations  - Structure the images in sensitive areas.  - Save the image in sequence.  of variations.  taken one after the other, until detection  - Compare zone by zone, the image recorded at SUi-  for the purpose of renovating the stored image.  - Then keep the affected areas of variations  stored image.  - Repeat the comparison cycle with the new one  2. Method according to claim 1, characterized by  a possibility of only raising awareness  areas for comparison allowing detection  of variation only on part of the Image,  3. Method according to claim 2, characterized by  fact that the part of the image made sensitive can  in addition constituted unboundings of image not necessarily pro  ches or adjacent.  a reduced set of these elements.  seems to be elements of these areas, or only on  that the zone comparison is carried out on the  4, Method according to claim 1, characterized in that  5. Process according to claim 1, characterized by  fact that the comparison may include an adjustment  threshold to raise awareness more or less the  detection of variations.  6. Method according to claim 1, characterized in that  that it can be used for any memory system  tion and transmission of images.  7. Method according to claim 1, characterized in that only the areas affected by a variation during the sequence of images can be stored or transmitted.  8. Method according to claim 1, characterized in that the variation detection makes it possible to memorize the image containing the cause of these variations.  9. Method according to claim 1, characterized by the possibility of using this processing with any technology allowing the storage and / or the transmission of sequences either of images, or of information other than those of images.  IO. Device for implementing the method according to claim 1, characterized by a transmitter-receiver unit connected to a telephone line, the transmitter of which performs the processing object of the invention and transmits the variations of successive zones to the receiver which updates its memory image at each arrival of zone information coming from the transmitter.