CN101472178B - Method and apparatus for encoding image - Google Patents

Abstract

The invention relates to a method and a device for encoding image. The method comprises the steps of starting, image acquisition, variable adjustment, encoding, storage and displaying. The device mainly comprises an image acquisition unit, an encode operation unit, a data storing unit, a display unit, wherein, the variable adjustment step is characterized in that image change rate and image complexity are utilized for the adjustment, the quantization scales of I-frame, P-frame, and B-frame, which are need by the encoding, are set, and the frame distance between the I-frame and the P-frame and the motion estimation search range of the I-frame and the P-frame, wherein, the image change rate is judged through the utilization of rates; and the image complexity is judged through the utilization of the entropic numerical value. Therefore, the image data is effectively compressed according to the running status, and the storage space is saved.

Description

Video encoding method and device

Technical field

The present invention relates to a kind of image coding mode and device.

Background technology

The generation of traffic accident, often human factor accounts for major part, and after traffic accident takes place, distinguishing traffic accident responsibility according to excuse, brake vestige or the Video Camera in street corner quite is not easy, also cause the traffic accident lawsuit continuous, more waste judicial resources, therefore there is the dealer that image acquisition system is installed on the vehicle, please refer to No. 095100761 " vehicle assisted monitoring apparatus and method " patent of the TaiWan, China patent No., it includes the following step: provide a detected light to detect the surrounding environment of a means of transportation outside; The reverberation that receives this detected light is to produce an image telecommunication signal; Carry out a sensor operation, produce a state signal with the state of this means of transportation of sensing; This state signal is carried out signal handle, to produce a controlling signal; And the catoptrical angle of adjusting the crevice projection angle and the intensity of this detected light and receiving this detected light according to this controlling signal;

And the TaiWan, China patent No. No. 094114989 " media recorder that is used for vehicle " also discloses similar patent case, comprising: whether a micro controller enters precarious position in order to judge vehicle, and receives the urgent signal in outer end; At least one image capture unit is in order to the input signal of acquisition at the beginning, and image output data; At least one voice capture device, in order to the input at the beginning the acquisition signal, and output speech data; One data processing core in order to handling the data that image capture unit and voice capture device are exported, and is compressed, is encrypted, and exported encrypt data; And at least one storing media, in order to the data of input data processing core output and store it, wherein, when the micro controller judgement has abnormal situation to occur, media recorder then starts, data processing core begins to receive the video-audio data from image capture unit and voice capture device, carries out the processing of data via data processing core, and the video-audio data after will handling then is stored to this storing media;

Other is such as No. 095201704, the TaiWan, China patent No. " automobile advance outside record and console screen image synchronous display system " patent, and the TaiWan, China patent No. 095223235 " vehicle image display system " patent all disclosed this type and utilized image capture unit, and method and device that institute's pick-up image is stored;

But existing automobile image acquisition equipment all focuses on the definition of image capture and how to use picked image analysis, when but all having ignored pick-up image, when the crossfire image that is captured stores, convert through coding, the image that then needs to store often needs a large amount of storage areas, the storage device that causes overflows easily, can't continue to store more images, can only store the preceding image continuation record that of short duration image frame then can't continue to use or select deletion at last, but which kind of mode no matter all can't be monitored, record work for a long time;

Though industrial circle has various existing coded system, but all lack and efficiently exercise road conditions at vehicle and encode, during particularly vehicle is advanced, road conditions difference is very big, the road surface scenery changed little when super expressway was sailed in for example general and words and deeds, the suitable great disparity of the peripheral picture difference in road surface when travelling the urban road, and be different from the definition of daytime running during night running, all multivariables like this all must be considered in detail;

Summary of the invention

Technical problem underlying to be solved by this invention is, overcome the above-mentioned defective that prior art exists, and provide a kind of video encoding method and device, can store more image datas, also can increase the time length of image capture, reach the record of best vehicle running state whereby, efficient according to travel conditions compressing image data, save the storage area.

Video encoding method of the present invention is:

A kind of video encoding method is characterized in that: mainly comprise the following steps:

Setting up procedure: when vehicle startup, system starts automatically, or manually boots system by the user, enters the image capture step after the startup;

Image capture step: utilize image acquisition unit, the road surface image data is captured, and the image that will capture gained is adjusted the variable step;

Adjust the parameter step: by the image rate of change, and the image complexity judge, adjust coding required I frame, P frame, B frame quantization scale, and set I frame, P frame frame spacing and move and estimate search area;

Coding step: the numerical value that sets according to above-mentioned adjustment variable step carries out image coding, and image data carries out storing step after will encoding, or carries out step display according to user's demand;

Storing step: the image data that will carry out behind the coding step is stored in data storage element;

Step display: the image data that will carry out behind the coding step is showed on the display unit;

By above-mentioned steps, with image capture step institute picked image data,, after adjusting, carry out coding step according to image rate of change and image complexity, reach the effect of efficient storage image data whereby.

Aforesaid video encoding method, wherein adjust in the variable step, the image rate of change is to utilize speed to judge, when speed is to be made as low image rate of change during less than 30km/hr, speed is between the 30km/hr to 80km/hr being moderate image rate of change, and speed is height image rate of change during greater than 80km/hr.

Aforesaid video encoding method is wherein adjusted in the variable step, and this image complexity is to utilize the numerical value of entropy to judge, the formula of this entropy is:

$\mathrm{\η}=H\left(S\right)=-\underset{i=1}{\overset{q}{\mathrm{\Σ}}}P\left(s_i\right){\log }_{2}P\left(s_i\right),$ S＝{s ₁，s ₂，Λ，s _q-1，s _q}；

S is the set that occurs symbol in the image data, and q is the sum of S set, and the symbol of Si for occurring, P (Si) then are the probability that corresponding symbol occurs;

When the numerical value of calculating entropy is to be made as low image complexity during less than 0.5/unit picture element, when the numerical value of entropy is to be to be made as moderate image complexity during between 0.5/unit picture element to 1.0/unit picture element, is to be to be made as height image complexity during greater than 1.0/unit picture element when the numerical value of entropy.

Aforesaid video encoding method, wherein adjust in the variable step, the image rate of change is to utilize speed to judge, when speed is made as low image rate of change during less than 30km/hr, speed is between the 30km/hr to 80km/hr being moderate image rate of change, and speed is height image rate of change during greater than 80km/hr; This image complexity is to utilize the numerical value of entropy to judge, the formula of this entropy is:

$\mathrm{\η}=H\left(S\right)=-\underset{i=1}{\overset{q}{\mathrm{\Σ}}}P\left(s_i\right){\log }_{2}P\left(s_i\right),$ S＝{S ₁，s ₂，Λ，s _q-1，s _q}；

S is the set that occurs symbol in the image data, and q is the sum of S set, and the symbol of Si for occurring, P (Si) then are the probability that corresponding symbol occurs; When the numerical value of calculating entropy is to be made as low image complexity during less than 0.5/unit picture element, when the numerical value of entropy is to be made as moderate image complexity between in 0.5/unit picture element to 1.0/unit picture element the time, when the numerical value of entropy is made as height image complexity during greater than 1.0/unit picture element;

The I of this minuent image rate of change and low image complexity, P frame spacing are 30: 15; The I of this minuent image rate of change and middle image complexity, P frame spacing are 30: 10; The I of this minuent image rate of change and high image complexity, P frame spacing are 30: 05; The I of this moderate image rate of change and low image complexity, P frame spacing are 20: 10; The I of this moderate image rate of change and middle image complexity, P frame spacing are 20: 05; The I of this moderate image rate of change and high image complexity, P frame spacing are 15: 05; The I of this height image rate of change and low image complexity, P frame spacing are 10: 05; The I of this height image rate of change and middle image complexity, P frame spacing are 06: 03; The I of this height image rate of change and high image complexity, P frame spacing are 06: 02.

Aforesaid video encoding method, wherein adjust in the variable step, the image rate of change utilizes speed to judge, when speed is made as low image rate of change during less than 30km/hr, speed is between the 30km/hr to 80km/hr being moderate image rate of change, and speed is height image rate of change during greater than 80km/hr; This image complexity utilizes the numerical value of entropy to judge, the formula of this entropy is:

$\mathrm{\η}=H\left(S\right)=-\underset{i=1}{\overset{q}{\mathrm{\Σ}}}P\left(s_i\right){\log }_{2}P\left(s_i\right),$ S＝{s ₁，s ₂，Λ，s _q-1，s _q}；

S is the set that occurs symbol in the image data, and q is the sum of S set, and the symbol of Si for occurring, P (Si) then are the probability that corresponding symbol occurs; When the numerical value of calculating entropy is to be made as low image complexity during less than 0.5/unit picture element, when the numerical value of entropy is made as moderate image complexity between in 0.5/unit picture element to 1.0/unit picture element the time, when the numerical value of entropy is made as height image complexity during greater than 1.0/unit picture element;

The mobile search area of estimating of this minuent image rate of change and low image complexity is 5; The mobile search area of estimating of this minuent image rate of change and middle image complexity is 5; The mobile search area of estimating of this minuent image rate of change and high image complexity is 5; The mobile search area of estimating of this moderate image rate of change and low image complexity is 10; The mobile search area of estimating of this moderate image rate of change and middle image complexity is 10; The mobile search area of estimating of this moderate image rate of change and high image complexity is 15; The mobile search area of estimating of this height image rate of change and low image complexity is 15; The mobile search area of estimating of this height image rate of change and middle image complexity is 15; The mobile search area of estimating of this height image rate of change and high image complexity is to be 15.

Image coding device of the present invention is:

A kind of image coding device is characterized in that: mainly comprise an image acquisition unit, coding arithmetic element, a data storage element and a display unit, wherein:

This image acquisition unit is to be installed in vehicle, utilizes this image acquisition unit to shoot with video-corder the raw video data, and with the raw video transfer of data of being shot with video-corder to encoding arithmetic element;

This coding arithmetic element is to receive the raw video data that image acquisition unit is shot with video-corder, adjust according to image rate of change and image complexity, set coding required I frame, P frame, B frame quantization scale, and set I frame, P frame frame spacing and move and estimate search area and encode, and the image data after will encoding is sent to data storage element or display unit;

This data storage element is that the image data behind the coding arithmetic element coding is stored, in order to use in the future;

This display unit is that the image data behind the coding arithmetic element coding is play, and watches for the user;

By the coding arithmetic element, make the image data of being shot with video-corder through after encoding, efficient compression stores the required capacity of putting that accounts for, and can reach best image data image quality.

Existing vehicle pick-up image device is only to capture and store at image data, not for various pavement behaviors, speed adjustment, and causes the storage area waste; The present invention utilizes image rate of change and image complexity, coded system is adjusted, high, normal, basic by rate setting image rate of change, and high, normal, basic with the setting value image complexity of entropy, change the quantization scale of the required I frame of coding, P frame, B frame according to image rate of change and image complexity, and set I frame, P frame frame spacing and move and estimate search area, so can efficiently make the more efficient compression of institute's picked image data; Image data by PSNR parameter checking computations after coded, in human eye sensitive scope do full blast compression and can keep higher image quality, if PSNR is less than 36dB, then adjust the variable step again, so through coding, can try to achieve optimum encoding parameter setting table after the checking computations, use this coding parameter setting table can guarantee that data storage element can effectively use, and be unlikely to waste, and can store more image datas, also can increase the time length of image capture, reach the record of best vehicle running state whereby, and can be efficient according to travel conditions compressing image data, the storage area saved.

Description of drawings

The present invention is further described below in conjunction with drawings and Examples.

Fig. 1 is a video encoding method steps flow chart schematic diagram of the present invention.

Fig. 2 is an image coding apparatus system configuration schematic diagram of the present invention.

Embodiment

As shown in Figure 1, the present invention is relevant a kind of " video encoding method and device ", and it mainly comprises the following steps: this video encoding method

Setting up procedure: when vehicle startup, system starts automatically, or manually boots system by the user, enters the image capture step after the startup;

Image capture step: utilize image acquisition unit, the road surface image data is captured, and the image that will capture gained is adjusted the variable step;

Adjust the parameter step: utilize image rate of change and image complexity to adjust, set coding required I frame, P frame, B frame quantization scale (I frame: Intra frame; P frame: Predicted frame; The B frame: Bi-directional frame), and set I frame, P frame frame spacing and move and estimate search area, wherein:

This image rate of change is to utilize speed to judge, when speed is made as low image rate of change during less than 30km/hr, speed is between the 30km/hr to 80km/hr being moderate image rate of change, and speed is height image rate of change during greater than 80km/hr;

This image complexity is to utilize the numerical value of entropy (Entropy) to judge that the formula of this entropy is:

$\mathrm{\η}=H\left(S\right)=-\underset{i=1}{\overset{q}{\mathrm{\Σ}}}P\left(s_i\right){\log }_{2}P\left(s_i\right),$ S＝{s ₁，s ₂，Λ，s _q-1，s _q}；

S is the set that occurs symbol in the image data, and q is the sum of S set, and the symbol of Si for occurring is if the object of hypothesis coding is a, b, four letters of c, d, S1=a then, S2=b, S3=c, S4=d, P (Si) then are the probability that corresponding symbol occurs, when the numerical value of calculating entropy less than 0.5

Image rate of change (speed)	Image complexity (numerical value of entropy)	I, P, B quantization scale	I, P frame spacing	Move and estimate search area
					Low	Low	(06/06/10)	(30/15)	5
Low	In	(08/08/10)	(30/10)	5
					Low	High	(08/10/10)	(30/05)	5
In	Low	(10/10/10)	(20/10)	10
					In	In	(10/12/12)	(20/05)	10
In	High	(12/14/14)	(15/05)	15
					High	Low	(15/15/15)	(10/05)	15
High	In	(20/16/16)	(06/03)	15
					High	High	(20/20/20)	(06/02)	15

Be made as low image complexity during/unit picture element (bit/per pixel), when the numerical value of entropy is made as moderate image complexity between in 0.5/unit picture element to 1.0/unit picture element the time, when the numerical value of entropy is made as height image complexity during greater than 1.0/unit picture element;

Please refer to table 1, table 1 is an image coding apparatus system allocation list of the present invention.

Table 1:

When low image rate of change and low image complexity, this I frame: P frame: B frame quantization scale ratio is to be 6: 6: 10, and I, P frame spacing are 30: 15, and moving and estimating search area is 5;

When low image rate of change and middle image complexity, this I frame: P frame: B frame quantization scale ratio is to be 8: 8: 10, and I, P frame spacing are 30: 10, and moving and estimating search area is 5;

When low image rate of change and high image complexity, this I frame: P frame: B frame quantization scale ratio is to be 8: 10: 10, and I, P frame spacing are 30: 5, and moving and estimating search area is 5;

When moderate image rate of change and low image complexity, this I frame: P frame: B frame quantization scale ratio is 10: 10: 10, and I, P frame spacing are 20: 10, and moving and estimating search area is 10;

When moderate image rate of change and middle image complexity, this I frame: P frame: B frame quantization scale ratio is 10: 12: 12, and I, P frame spacing are 20: 5, and moving and estimating search area is 10;

When moderate image rate of change and high image complexity, this I frame: P frame: B frame quantization scale ratio is 12: 14: 14, and I, P frame spacing are 15: 5, and moving and estimating search area is 15;

When height image rate of change and low image complexity, this I frame: P frame: B frame quantization scale ratio is 15: 15: 15, and I, P frame spacing are 10: 5, and moving and estimating search area is 15;

When height image rate of change and middle image complexity, this I frame: P frame: B frame quantization scale ratio is to be 20: 16: 16, and I, P frame spacing are 6: 3, and moving and estimating search area is to be 15;

When height image rate of change and high image complexity, this I frame: P frame: B frame quantization scale ratio is to be 20: 20: 20, and I, P frame spacing are 6: 2, and moving and estimating search area is to be 15;

Training step: through the image data of coding back gained is to utilize the calculating of testing of following formula, then do not adjust the parameter step again if meet: utilize PSNR (PeakSignal-to-Noise Ration, the signal peak noise ratio) parameter checks, and it is defined as follows:

Wherein MES (Mean Squared Error) is defined as:

I is the width index value of image data, and j is the height index value of image data, and f _{I, j}With f ' _{I, j}Represent respectively after unpressed image frame and the compressed decoding image frame (i, j) the picture element value of position, M be image data width (unit: picture element), N be image data height (unit: picture element), and 1≤i≤M, 1≤j≤N; Because the mankind can differentiate the good corrupt limited of image data picture, when PSNR value during more than or equal to 36dB, the sensible image data picture of naked eyes institute is clear, therefore the PSNR value that is checked is if be lower than 36dB, then must carry out aforesaid adjustment parameter step again, in the hope of reaching optimal compression ratio state;

Coding step: the numerical value that sets according to above-mentioned adjustment variable step carries out image coding, and image data carries out storing step after will encoding, or carries out step display according to user's demand;

Storing step: the image data that will carry out behind the coding step is stored in data storage element;

Step display: the image data that will carry out behind the coding step is showed on the display unit;

By above-mentioned steps, with image capture step institute picked image data, according to image rate of change and image complexity, through adjusting I frame, P frame, B frame quantization scale, set I frame, P frame frame spacing and the mobile search area of estimating, so that carry out coding step, reach the effect of efficient storage image data whereby.

As shown in Figure 2, image coding device of the present invention, it mainly comprises an image acquisition unit, coding arithmetic element, a data storage element and a display unit, wherein:

This image acquisition unit is to be installed in vehicle, utilizes this image acquisition unit to shoot with video-corder the raw video data, and with the raw video transfer of data of being shot with video-corder to encoding arithmetic element;

This coding arithmetic element is to receive the raw video data that image acquisition unit is shot with video-corder, according to image rate of change, required adjustment I frame, P frame, the B quantization scale of image complexity adjustment coding, set I frame, P frame frame spacing and the mobile search area of estimating, the raw video data are encoded, and the image data after will encoding is sent to data storage element or display unit;

This data storage element is that the image data behind the coding arithmetic element coding is stored, in order to use in the future;

This display unit is that the image data behind the coding arithmetic element coding is play, and watches for the user;

By the coding arithmetic element, make the image data of being shot with video-corder through after encoding, efficient compression stores the required capacity of putting that accounts for, and can reach best image data image quality.

Claims (6)

1. a video encoding method is characterized in that: mainly comprise the following steps:

Setting up procedure: when vehicle startup, system starts automatically, or manually boots system by the user, enters the image capture step after the startup;

Image capture step: utilize image acquisition unit, the road surface image data is captured, and the image that will capture gained is adjusted the variable step;

Adjust the parameter step: by the image rate of change, and the image complexity judge, adjust coding required I frame, P frame, B frame quantization scale, and set I frame, P frame frame spacing and move and estimate search area;

Coding step: the numerical value that sets according to above-mentioned adjustment variable step carries out image coding, and image data carries out storing step after will encoding, or carries out step display according to user's demand;

Storing step: the image data that will carry out behind the coding step is stored in data storage element;

Step display: the image data that will carry out behind the coding step is showed on the display unit;

By above-mentioned steps, with image capture step institute picked image data,, after adjusting, carry out coding step according to image rate of change and image complexity, reach the effect of efficient storage image data whereby.

2. video encoding method according to claim 1, it is characterized in that: in the described adjustment variable step, the image rate of change is to utilize speed to judge, when speed is to be made as low image rate of change during less than 30km/hr, speed is between the 30km/hr to 80km/hr being moderate image rate of change, and speed is height image rate of change during greater than 80km/hr.

3. video encoding method according to claim 1 is characterized in that: in the described adjustment variable step, this image complexity is to utilize the numerical value of entropy to judge, the formula of this entropy is:

$\mathrm{\η}=H\left(S\right)=-\underset{i=1}{\overset{q}{\mathrm{\Σ}}}P\left(s_i\right){\log }_{2}P\left(s_i\right),$ S＝{s ₁，s ₂，Λ，sq _-1，s _q}；

S is the set that occurs symbol in the image data, and q is the sum of S set, and the symbol of Si for occurring, P (Si) then are the probability that corresponding symbol occurs;

When the numerical value of calculating entropy is to be made as low image complexity during less than 0.5/unit picture element, when the numerical value of entropy is to be to be made as moderate image complexity during between 0.5/unit picture element to 1.0/unit picture element, is to be to be made as height image complexity during greater than 1.0/unit picture element when the numerical value of entropy.

4. video encoding method according to claim 1, it is characterized in that: in the described adjustment variable step, the image rate of change is to utilize speed to judge, when speed is made as low image rate of change during less than 30km/hr, speed is between the 30km/hr to 80km/hr being moderate image rate of change, and speed is height image rate of change during greater than 80km/hr; This image complexity is to utilize the numerical value of entropy to judge, the formula of this entropy is:

$\mathrm{\η}=H\left(S\right)=-\underset{i=1}{\overset{q}{\mathrm{\Σ}}}P\left(s_i\right){\log }_{2}P\left(s_i\right),$ S＝{s ₁，s ₂，Λ，s _q-1，s _q}；

S is the set that occurs symbol in the image data, and q is the sum of S set, and the symbol of Si for occurring, P (Si) then are the probability that corresponding symbol occurs; When the numerical value of calculating entropy is to be made as low image complexity during less than 0.5/unit picture element, when the numerical value of entropy is to be made as moderate image complexity between in 0.5/unit picture element to 1.0/unit picture element the time, when the numerical value of entropy is made as height image complexity during greater than 1.0/unit picture element;

The I of this minuent image rate of change and low image complexity, P frame spacing are 30:15; The I of this minuent image rate of change and middle image complexity, P frame spacing are 30:10; The I of this minuent image rate of change and high image complexity, P frame spacing are 30:05; The I of this moderate image rate of change and low image complexity, P frame spacing are 20:10; The I of this moderate image rate of change and middle image complexity, P frame spacing are 20:05; The I of this moderate image rate of change and high image complexity, P frame spacing are 15:05; The I of this height image rate of change and low image complexity, P frame spacing are 10:05; The I of this height image rate of change and middle image complexity, P frame spacing are 06:03; The I of this height image rate of change and high image complexity, P frame spacing are 06:02.

5. video encoding method according to claim 1, it is characterized in that: in the described adjustment variable step, the image rate of change utilizes speed to judge, when speed is made as low image rate of change during less than 30km/hr, speed is between the 30km/hr to 80km/hr being moderate image rate of change, and speed is height image rate of change during greater than 80km/hr; This image complexity utilizes the numerical value of entropy to judge, the formula of this entropy is:

$\mathrm{\η}=H\left(S\right)=-\underset{i=1}{\overset{q}{\mathrm{\Σ}}}P\left(s_i\right){\log }_{2}P\left(s_i\right),$ S＝{s ₁，s ₂，Λ，s _q-1，s _q}；

S is the set that occurs symbol in the image data, and q is the sum of S set, and the symbol of Si for occurring, P (Si) then are the probability that corresponding symbol occurs; When the numerical value of calculating entropy is to be made as low image complexity during less than 0.5/unit picture element, when the numerical value of entropy is made as moderate image complexity between in 0.5/unit picture element to 1.0/unit picture element the time, when the numerical value of entropy is made as height image complexity during greater than 1.0/unit picture element;

The mobile search area of estimating of this minuent image rate of change and low image complexity is 5; The mobile search area of estimating of this minuent image rate of change and middle image complexity is 5; The mobile search area of estimating of this minuent image rate of change and high image complexity is 5; The mobile search area of estimating of this moderate image rate of change and low image complexity is 10; The mobile search area of estimating of this moderate image rate of change and middle image complexity is 10; The mobile search area of estimating of this moderate image rate of change and high image complexity is 15; The mobile search area of estimating of this height image rate of change and low image complexity is 15; The mobile search area of estimating of this height image rate of change and middle image complexity is 15; The mobile search area of estimating of this height image rate of change and high image complexity is to be 15.

6. image coding device is characterized in that: mainly comprise an image acquisition unit, coding arithmetic element, a data storage element and a display unit, wherein:

This image acquisition unit is to be installed in vehicle, utilizes this image acquisition unit to shoot with video-corder the raw video data, and with the raw video transfer of data of being shot with video-corder to encoding arithmetic element;

This coding arithmetic element is to receive the raw video data that image acquisition unit is shot with video-corder, adjust according to image rate of change and image complexity, set coding required I frame, P frame, B frame quantization scale, and set I frame, P frame frame spacing and move and estimate search area and encode, and the image data after will encoding is sent to data storage element or display unit;

This data storage element is that the image data behind the coding arithmetic element coding is stored, in order to use in the future;

This display unit is that the image data behind the coding arithmetic element coding is play, and watches for the user.

Images