KR100331709B1 - Vehicle image data compression system using a wavelet - Google Patents

Abstract

The present invention relates to an apparatus for compressing vehicle image information using wavelets, and in the related art, it is only mentioned that the vehicle image data obtained is stored in a storage device, but it is not known how to make and store the vehicle image data in an optimal state. There is a problem of not knowing how to transmit the digital image data to a remote location. Therefore, the present invention captures the vehicles driving on the road, and outputs the digital vehicle image data taken by the camera, the digital vehicle image data photographed through the camera is input to extract the traffic information when the event occurs by the program, Problems caused by a vehicle detection computer for determining whether to convert wavelets according to a result of checking whether a predefined problem has occurred, and traffic information extracted from digital vehicle image data captured by the camera according to the result determined by the vehicle detection computer. A wavelet video encoder including an ID and compressed with a wavelet transform, a memory for storing the compressed video information data compressed above, and compressed video information data compressed by the wavelet video encoder or a block stored in the memory. A container for sending data to a remote location After the composed channel, compressed without loss of information as a wavelet transform is one to be transferred to a remote location over a communication channel.

Description

Vehicle image information compression device using wavelet {VEHICLE IMAGE DATA COMPRESSION SYSTEM USING A WAVELET}

The present invention relates to a vehicle image information compression apparatus using a wavelet that is capable of storing or transmitting the image information of the vehicle sensed through a camera without loss by wavelet transformation, in particular, DCT (Discrete Coefficient Transform) The present invention relates to a vehicle image information compression apparatus using wavelets that can be compressed to high quality image data without blocking and corona phenomena, which are JPEG disadvantages.

In general, in the transportation system field, a camera is a sensor that monitors a vehicle and detects the flow of a vehicle. The vehicle license plate is combined with an image processing device to acquire a vehicle flow on a road and passes through a toll gate. It is a useful tool for remote monitoring.

In addition, the extraction of vehicle information can be detected using sensors such as loops, ultrasonic waves, infrared rays, microwaves, lasers, etc., but the vehicle is detected using a camera, and the current road driving state is indicated by the flow, queue, and occupancy rate of the vehicle. It can be seen.

There is a need to transmit the real image data of the traffic information as described above according to a remote command. In this case, the real image data of the traffic information is compressed and transmitted.

There are two methods of compressing digital image data: DCT and wavelet compression.

JPEG, MPEG, etc. can be implemented using the DCT method, but when using the DCT method, it is difficult to avoid a blocking effect or a corona noise effect.

However, if the wavelet compression method is used, high quality image compression is possible without loss, and compression is possible without blocking or corona phenomenon.

Patents applying the wavelet compression method are U.S. Pat. No. 5,754,793 and U.S. Pat. No. There is 5.066,950.

U.S. Pat. No. 5,754,793 use wavelet image compression and reconstruction device to model human visual system. Image quantization is extracted by wavelet transform.

U.S. Pat. No. 5.066, 950 stores image data recording media in a video cassette (VCR), and is applied to a vehicle monitoring system.

The vehicle monitoring apparatus using the wavelet compression method as described above will be described with reference to FIG. 1.

1 is a configuration diagram of a vehicle monitoring apparatus using a conventional wavelet, as shown in FIG. 1, a pair of infrared sensors 1 installed at a predetermined interval for determining whether the vehicle is entered or not by the detected infrared rays, and A CD camera 3 for acquiring an image of the vehicle on the road according to a control signal and the pair of infrared sensors 1 are notified whether the vehicle is entering or leaving the vehicle, and the CD camera is in accordance with a program command. The microcomputer 2 determines whether to operate (3) and controls the function of storing or recording the image data received through the CD camera 3, and the control of the microcomputer 2 Prints the VCD 4 for recording the image data acquired through the CD camera 3 and the image data acquired through the CD camera 3 under the control of the microcomputer 2. Is composed of a printer 5, a monitor 6 which monitors the video obtained through the ssissi di camera 3 is under the control of the microcomputer (2).

Looking at the prior art configured in this way in detail as follows.

A pair of infrared sensors 1 provided at regular intervals detect the infrared rays generated by the infrared light emitting unit, and provide the detection signal to the microcomputer 2.

Then, the microcomputer 2 determines the entry and exit of the vehicle according to the infrared detection signal provided from the infrared sensor 1.

That is, when an infrared ray detected signal is input, it is determined that there is no entry of the vehicle, and when a signal not detected infrared ray is input, it is determined that there is an entry of the vehicle.

For example, when it is determined that the vehicle has entered, the microcomputer 2 operates a program stored therein to generate a command to shoot with the CD camera 3, and the video data to be captured by the VCD 4. Generates a control signal to save.

The CD camera 3 receiving the command generated by the microcomputer 2 acquires an image of a vehicle passing through a road, and provides the obtained image data to the microcomputer 2.

The microcomputer 2 calculates the speed, the headway and the length of the vehicle from the acquired image data, and checks the consistency of the vehicle.

At the same time, when monitoring, the microcomputer 2 transmits the image data photographed through the CD camera 3 to the monitor 6 to enable display.

The monitor agent then monitors through the image displayed on the monitor 6.

The microcomputer 2 which checks the calculation and the density of the vehicle determines whether or not the speed of the calculated vehicle violates or violates using statistical file data stored in the internal memory.

When the data determined in this way is to be recorded, it is stored in the VRC (4).

When the data stored in the VRC 4 is to be printed, the microcomputer 2 controls the printer 5 to read the data from the VRL 4 and print the data.

When the monitor is to be monitored, the microcomputer 2 provides the monitor 6 with data read from the CD camera 3 and displays the data.

The operation so far is briefly described based on the operation flowchart shown in FIG. 2. When an event (EVENT) such as an entry of a vehicle occurs using infrared rays detected by the infrared sensor 1 (S11), the microcomputer ( 2) controls the CD camera 3 to acquire the image data for the vehicle, (S12)

Using the image data thus obtained, the microcomputer 2 calculates the speed, the distance between vehicles, and the length of the entered vehicle (S13), and checks the vehicle density based on the calculated result (S14).

The result checked in step S14 is compared with the data stored in the statistics file of the internal memory to check whether the currently detected vehicle is violated.

The result of this check is recorded in a storage device such as VRT (S16).

However, in the prior art as described above, it is only mentioned that the image data acquired through the CD camera is stored in a storage device such as VCD, but it is not known how to make and store the digital image data in an optimal state. There is a problem of not knowing how to transmit the digital image data to a remote location.

Accordingly, an object of the present invention for solving the conventional problems as described above is to provide a vehicle image information compression apparatus using a wavelet that is compressible to high quality image data without blocking and corona phenomenon, which is a disadvantage of JPEG compression by DCT method. .

Another object of the present invention is to provide a vehicle image information compression apparatus using a wavelet that can be compressed without loss of information by wavelet transformation.

It is still another object of the present invention to provide a vehicle image information compression apparatus using wavelets for transmitting image data compressed by wavelet transformation to a remote place through a communication channel.

1 is a block diagram of a vehicle monitoring apparatus using a conventional wavelet.

2 is a flowchart illustrating an operation of a vehicle monitoring apparatus using a conventional wavelet.

3 is a block diagram of a vehicle image information compression apparatus using a wavelet of the present invention.

4 is a flowchart illustrating an operation state of the apparatus for compressing vehicle image information using wavelets according to the present invention.

5 is a detailed block diagram of the wavelet video encoder in FIG. 3; FIG.

FIG. 6 is a detailed block diagram of the wavelet kernel filter in FIG. 5; FIG.

7 is a detailed block diagram of the adaptive quantizer in FIG.

***** Explanation of symbols for main parts of drawing *****

DESCRIPTION OF SYMBOLS 100 Camera 200 Vehicle detection computer 201 Wavelet kernel filter 202 Adaptive quantizer

203 run-length and Huffman encoder

300: wavelet video encoder 400: flash memory

500: communication channel 600: monitor

The present invention for achieving the above object is a camera for photographing the driving state of the road and providing the digital image data photographed, traffic information when the event occurs by the program, and determines whether to compress the wavelet, compressed image information The vehicle detection computer performing an operation for storing and transmitting the data, and if the image data is required to be stored in case of an event, the image data including a problem occurrence ID when the problem occurs in advance according to the traffic information is compressed by wavelet transformation and then the vehicle is compressed. Compression means provided to the detection computer, storage means for storing the compressed image information data compressed by the compression means, and communication means for transmitting the compressed image information data compressed by the compression means to a remote location.

Hereinafter, with reference to the accompanying drawings in detail as follows.

3 is a block diagram of a vehicle image information compression apparatus using wavelets according to the present invention. As shown in FIG. 3, a camera 100 which photographs a driving state of a road and outputs the photographed digital vehicle image data; A vehicle detection computer that receives the digital vehicle image data output from the camera 100 and performs vehicle detection, determines whether to compress the wavelet by a program, and performs a control operation for storing and transmitting the compressed image data. 200 and a wavelet video encoder 300 including a problem occurrence ID pre-defined by traffic information under the control of the vehicle detection computer 200 in the video data of the camera 100 and compressing the wavelet by compression. And the image data compressed by the wavelet video encoder 300 through wavelet transformation to control of the vehicle detection computer 200. And a communication channel 500 for transmitting the image data compressed by the wavelet video encoder 300 to a remote location under the control of the vehicle detection computer 200 and the camera ( The monitor 600 is configured to display the digital vehicle image data transmitted through the display 100 on a screen to enable monitoring.

As shown in FIG. 5, the wavelet video encoder 300 includes a wavelet kernel filter 201 for performing high pass and low pass filtering on the input digital vehicle image data, sampling, and reducing data. An adaptive quantizer 202 for quantizing data output from the wavelet kernel filter 201, and a run-length code and Huffman coding for compressing the quantized data by the adaptive quantizer 202. It consists of the length and Huffman encoder 203.

As shown in FIG. 6, the wavelet kernel filter 201 includes a low frequency image acquisition unit 301 for high frequency filtering and sampling low frequency image data of the vehicle image data, and the vehicle. Low-frequency region acquisition unit 302 for low-pass filtering and sampling image data to obtain high-frequency image data, and low-frequency region for subdividing and compressing low-frequency image data acquired by the low-frequency image acquisition unit 301. And an installment unit 303 and a high frequency region dividing unit 304 for dividing and compressing the high frequency image data acquired by the high frequency image acquisition unit 302.

As shown in FIG. 7, the adaptive quantizer 202 includes a band coefficient generator 402 for generating a band coefficient having a minimum distortion, kernel filtered image data, and generation of the band coefficient. A multiplier 401 multiplying the band coefficient generated by the unit 402, an adder 403 for adding a quantization coefficient to a value obtained through the multiplier 401, and the remaining digits from the value added by the adder 403 is removed. The remaining digit removing unit 404 for outputting the obtained quantized value is configured.

When described in detail with respect to the operation and effect of the present invention configured as described above.

When digital image data is compressed by DCT conversion, pixel values that are irregularly spread on the screen before conversion tend to be concentrated toward the low frequency term after conversion.

Therefore, the high frequency terms are discarded after the DCT conversion, so that the information is compressed with little information loss.

However, when a block is divided into 8 × 8 small blocks and compressed using a DCT transform, a block-to-block blur occurs.

Therefore, in the present invention, the image data is compressed into a high frequency region pass filter and a low frequency region pass filter by a wavelet transform in which there is almost no loss of information in a given frequency space through a continuous dividing operation in the horizontal direction and the vertical direction.

In addition, by transmitting the wavelet coefficients used in the wavelet transform and using the wavelet coefficients received during wavelet recovery, errors in the inter-frame breaking sequence are reduced, and the original time series Enable data recovery.

Referring to such an operation, the camera 100 of FIG. 3 continuously photographs the driving state of the road, and provides digital vehicle image data obtained by the photographing to the vehicle detection computer 200 and the wavelet video encoder 300, respectively. do.

Then, the vehicle detection computer 200 receives the digital vehicle image data provided from the camera 100 and determines whether an event such as vehicle entry occurs by a program (S101).

As a result of the determination, if the event does not occur, it is continuously determined whether the event occurs, and when the event occurs, the queue, occupancy rate, speed, etc. of the vehicle are calculated to extract traffic information.

In addition, the vehicle detection computer 200 provides the wavelet video encoder 300 with a problem ID according to traffic information when the image data needs to be stored when an event occurs or when a predefined critical issue occurs.

Then, the wavelet video encoder 300 includes a problem ID by the traffic information provided by the vehicle detection computer 200 in the digital vehicle image data provided by the camera 100 to perform wavelet conversion and compress it. (S103)

The compressed data is compressed video information data.

The compressed video information data compressed by the wavelet video encoder 300 is stored in the flash memory 400 in the case of simple storage, and transmitted to the monitor 600 for monitoring.

When real-time transmission of the compressed video information data is required, it is transmitted to the communication channel 500 (S104).

As such, operations of the flash memory 400, the communication channel 500, and the monitor 600 are controlled by the vehicle detection computer 200.

In addition, the vehicle detection computer 200 reads the block data stored in the flash memory 400 to the communication channel 500 when an event for transmitting the block data already stored in the flash memory 400 to a remote location occurs. do.

The communication channel 500 then transmits block data.

The wavelet video encoder 300 operating as described above will be described in more detail with reference to FIG. 5, wherein the vehicle detection computer is stored in the digital vehicle image data or the digital vehicle image data from the camera 100 photographing the driving state of the road. The wavelet kernel filter 201 receives image information data including the traffic information or the problem occurrence ID provided by the 200.

The wavelet kernel filter 201 receiving the image information data first performs the high pass filtering and the low pass filter, and then divides the image through several times by reducing the sampling by ½ in the horizontal direction and the vertical direction, respectively. Obtain the image data value.

A detailed operation of the wavelet kernel filter 201 for obtaining image data values for dividing the screen will be described later.

In the case of the image, if the contrast of the object is relatively low, it is close to the low frequency band, and the edge of the object with the sharp change of black and white contrast is said to be close to the high frequency band.

In general, if data is compressed by decomposing image data in a high frequency region into a low frequency region like a human visual system, loss of information becomes a problem.

In order to solve this problem, the adaptive quantizer 202 receives image data of the divided image through the wavelet kernel filter 201 and quantizes the image data according to the information contained in the bandpass filtered image. Quantize the quality without damage.

Herein, the operation of the adaptive quantizer 202 will be described based on FIG. 7. The data output from the wavelet kernel filter 201 is received by the multiplier 401 as one input terminal.

In this case, the band coefficient generator 402 refers to a coefficient having a minimum distortion different from each wavelet block as a band coefficient, and generates the band coefficient.

Then, the multiplier 401 receives the band coefficient as another input terminal and multiplies the band coefficient by the data output from the wavelet kernel filter 201.

When the multiplier 403 is input to the multiplier 403 and the quantization coefficient is added to the remaining digit removing unit 404, the remaining digit removing unit 404 outputs a value in which the remaining digits except the quotient are removed. This value is the quantization value.

In this way, the quantization value is inputted from the run-length and Huffman encoder 203, and a short length code is assigned to a non-zero coefficient that has a high probability of occurrence, that is, a RUN of zero coefficients, and a low probability of occurrence. A long code is assigned to the coefficient.

Changing the coding scheme in this way can reduce the average code length.

The value obtained by the run-length encoding is compared using a Huffman table to obtain a value corresponding to the table value. The obtained value is compressed data. That is, it becomes compressed video information data.

Until now, the wavelet kernel filter 201, the adaptive quantizer 202, the run-length and the Huffman coder 203 are sequentially encoded.

The restoration of the image data may be performed in the reverse order.

Referring to the operation of the wavelet kernel filter 201 based on FIG. 6, first, image information data (f (x, y) including traffic information and problem IDs are included in the digital vehicle image data from the camera 100. When the)) is input, the low frequency image acquisition unit 301 performs high pass filtering in the high pass filter 11, and performs sampling by reducing the filtered data by the sampling unit 12 to ½ in the horizontal direction (x direction). The low frequency image data f _L (x, y) is obtained, the high frequency image acquisition unit 302 performs low pass filtering by the low pass filter 21, and the low pass filtered data by the sampling unit 22 in the horizontal direction ( Sampling is performed in half in the x direction) to obtain high frequency image data f _H (x, y).

The low frequency image data f _L (x, y) thus obtained is provided to the first frequency domain divider 303, and the high frequency image data f _H (x, y) is provided to the second frequency domain divider 304. To provide.

Accordingly, the first frequency domain divider 303 again filters the low frequency image data f _L (x, y) with the high pass filter 13 and the low pass filter 15, respectively, and then the sampling unit 14 ( In 16), the sample is reduced by ½ in the vertical direction (y direction) to divide the low frequency region into an upper region f _LH (x, y) and a lower region f _LL (x, y).

Similarly, the second frequency domain divider 304 filters the high frequency image data f _H (x, y) with the high pass filter 23 and the low pass filter 25, respectively, and then the sampling unit 24 ( In step 26), the high frequency region is divided into the upper region f _HL (x, y) and the lower region f _HH (x, y) by sampling ½ in the vertical direction (y direction).

As a result, the image data in the high frequency region is decomposed to the resolution of the low frequency region and then quantized so as not to degrade the visual quality.

According to this compression method, if the image data of 512 × 512 pixels, that is, the resolution of one pixel is 8 bits, the information amount of one screen of 512 × 512 × 8 is 256 kbytes of data. When the compression ratio is adjusted (¼ to 1/3360), compression is possible from 64 Kbytes to 0.8 Kbytes, and the compressed digital video data is stored in a memory resistant to environmental and vibration resistance suitable for outdoor environments. You can enable storage.

In addition, such a vehicle image compression device can be used in industrial numbers, character recognition, image detectors, unmanned surveillance, satellite image analysis and military image processing applications.

As described in detail above, the present invention extracts traffic information when recognizing a vehicle, compresses and converts only the problem image data to 1 / 4-1 / 350 in a flash memory, and stores it in a flash memory, using a communication means to a remote location. It has the effect of being able to transfer.

Claims (5)

Photographing means for photographing vehicles traveling on the road, and outputting the photographed digital vehicle image data; and extracting traffic information when an event occurs by a program by receiving digital vehicle image data photographed through the photographing means, and extracting the extracted information. A vehicle detection computer for determining whether or not to convert a wavelet according to a result of checking a predefined problem from the traffic information, and traffic extracted from digital vehicle image data photographed through the photographing means according to the result determined by the vehicle detection computer. Compression means for compressing by wavelet transformation including a problem occurrence ID caused by the information, Storage means for storing the compressed video information data compressed above, Compressed video information data compressed by the compression means or stored in the storage means Communication means for transmitting the block data Vehicle image information compression apparatus using a wavelet, characterized in that configured to include. The wavelet kernel filtering unit of claim 1, wherein the compression unit performs high-pass and low-pass filtering on the input digital vehicle image data, cuts the filtered data in half, and divides the sampled image into smaller pieces, and the wavelet. Quantization means for quantizing the divided data output through the kernel filtering means without loss of visual quality, and after performing run-length coding on the quantized data by the quantization means, a compressed code value corresponding to the encoded value is Huffman table. Vehicle image information compression apparatus using a wavelet, characterized in that it comprises an encoding means found in. 3. The apparatus of claim 2, wherein the wavelet kernel filtering means performs a high pass filtering of the input image data, compresses the image data in a horizontal direction, and performs sampling to obtain a low frequency image by performing sampling and low pass filtering of the image data. High frequency image acquisition means for acquiring a high frequency image by compressing and sampling in a horizontal direction, low frequency region dividing means for dividing the low frequency image data acquired by said low frequency image acquisition means and compressing it in a vertical direction, and said high frequency image And a high frequency region dividing means for dividing the high frequency image data acquired by the acquiring means and compressing the image data in the vertical direction. 4. The apparatus of claim 3, wherein the low frequency region dividing means comprises: high pass filtering means for high frequency filtering low frequency image data to obtain an upper region of the low frequency, and low pass filtering means for low frequency filtering the low frequency image data to obtain a lower region of the low frequency And sampling means for compressing and sampling the data output from the high pass and low pass filtering means in a vertical direction in half of the vehicle image information compression device using a wavelet. 4. The apparatus of claim 3, wherein the high frequency region dividing means is configured in the same manner as the low frequency region dividing means.