CN104767912A - Video defogging method and system based on FPGA - Google Patents

Abstract

The invention relates to a video defogging method based on an FPGA. With the FPGA as a processing body, through detecting the maximum value and the minimum value of R, G and B values of each pixel in a current video image frame, a maximum value matrix and a minimum value matrix are obtained, the average value of all the minimum values in the minimum value matrix is calculated, and average filtering is performed on the minimum value matrix; according to an obtained average filtering matrix and the obtained average value, an atmospheric transmission ratio matrix is calculated; according to the maximum value matrix, the average filtering matrix and the current video image frame, the light component of atmosphere is calculated; when a next video image frame is received, RGB data of the current video image frame are read and converted into YUV spatial data; according to the atmospheric transmission ratio matrix, the light component of the atmosphere and a Y value in the YUV spatial data, a Y value in the defogged YUV spatial data is calculated, wherein the Y value and a U value and a V value in the YUV spatial data are combined to form the defogged YUV spatial data; the defogged YUV spatial data are converted into RGB data, and the RGB data after defogging the current video image frame are obtained.

Description

Based on video defogging method capable and the system of FPGA

Technical field

The present invention relates to technical field of video image processing, particularly relate to a kind of video defogging method capable based on FPGA and system.

Background technology

The video taken under the weather condition having mist or haze, because the suspended particulate in air produces absorption and scattering process to light, cause image quality decrease, occur ambiguous phenomenon, reduce the visuality of video, and interference is caused to successive image process.Existing mist elimination algorithm speed is slow, and amount of calculation is large, and can not realize on hardware.

Summary of the invention

Technical problem to be solved by this invention is, proposes a kind of video defogging method capable based on FPGA and system, and slow to solve existing algorithm speed, amount of calculation is large, the difficult problem that can not realize on hardware.The present invention is achieved in that

Based on a video defogging method capable of FPGA, described method carries out mist elimination by FPGA to current frame video image, comprises the steps:

Receive and store current frame video image, meanwhile, detecting the maximum in R, G, B value of each pixel in described current frame video image and minimum value, obtaining maximum matrix and minimum value matrix;

Mean filter is carried out to described minimum value matrix, exports the mean filter matrix of minimum value;

Calculate the mean value of each minimum value in described minimum value matrix;

According to described mean filter matrix and described mean value calculation atmospheric transmission rate matrix;

Atmosphere light composition is calculated according to described maximum matrix, mean filter matrix and described current frame video image;

When receiving next frame video image, read the RGB data of described current frame video image, and be converted into yuv space data;

The Y value in the yuv space data after mist elimination is calculated according to the Y value in described atmospheric transmission rate matrix and atmosphere light composition and described yuv space data;

Yuv space data after utilizing the U value in the Y value in the yuv space data after mist elimination and described yuv space data and V value to form mist elimination;

Yuv space data after described mist elimination are converted to RGB data, obtain the RGB data after described current frame video image mist elimination.

Further, described current frame video image is stored in DDR memory.

Further, the formula calculating atmospheric transmission rate matrix is: L ₀=min (min (pm _av, 1.0) and M _ave(x), M (x)); Wherein, p is mist elimination intensity, m _avfor the mean value of each minimum value in described minimum value matrix; M _avex () is described mean filter matrix; M (x) is described current frame video image; M _ave(x)=average _sa(Min (x)), wherein, described Min (x) is described minimum value matrix.

Further, described method also comprises the steps:

PC regulates described FPGA by serial ports, to regulate the value of described p.

Further, the computing formula of described atmosphere light composition is: $A=\frac{1}{2}(\max \left(\mathrm{Max}\left(x\right)\right)+\max \mathrm{Mave}\left(x\right))\left[\begin{array}{c}1\\ 1\\ 1\end{array}\right].$

Based on a video mist elimination system of FPGA, described system carries out mist elimination by FPGA to current frame video image, comprising:

Matrix builds module, in order to receive and to store current frame video image, meanwhile, detects the maximum in R, G, B value of each pixel in described current frame video image and minimum value, obtains maximum matrix and minimum value matrix;

Mean filter module, in order to carry out mean filter to described minimum value matrix, exports the mean filter matrix of minimum value;

Mean value calculation module, in order to calculate the mean value of each minimum value in described minimum value matrix;

Atmospheric transmissivity matrix computations module, in order to according to described mean filter matrix and described mean value calculation atmospheric transmission rate matrix;

Atmosphere light composition computing module, in order to calculate atmosphere light composition according to described maximum matrix, mean filter matrix and described current frame video image;

Yuv space data conversion module, in order to when receiving next frame video image, reading the RGB data of described current frame video image, and being converted into yuv space data;

Y value computing module after mist elimination, in order to calculate the Y value in the yuv space data after mist elimination according to the Y value in described atmospheric transmission rate matrix and atmosphere light composition and described yuv space data;

Yuv space data construct module after mist elimination, forms the yuv space data after mist elimination in order to utilize the U value in the Y value in the yuv space data after mist elimination and described yuv space data and V value;

RGB data modular converter, in order to the yuv space data after described mist elimination are converted to RGB data, obtains the RGB data after described current frame video image mist elimination.

Further, described current frame video image is stored in DDR memory.

Further, the formula calculating atmospheric transmission rate matrix is: L ₀=min (min (pm _av, 1.0) and M _ave(x), M (x)); Wherein, p is mist elimination intensity, m _avfor the mean value of each minimum value in described minimum value matrix; M _avex () is described mean filter matrix; M (x) is described current frame video image; M _ave(x)=average _sa(Min (x)), wherein, described Min (x) is described minimum value matrix.

Further, described system also comprises PC, and described PC regulates described FPGA by serial ports, to regulate the value of described p.

Further, the computing formula of described atmosphere light composition is: $A=\frac{1}{2}(\max \left(\mathrm{Max}\left(x\right)\right)+\max \mathrm{Mave}\left(x\right))\left[\begin{array}{c}1\\ 1\\ 1\end{array}\right].$

Compared with prior art, the present invention, using FPGA as process main body, only needs storage one frame video, just can meet the function of fast in real time mist elimination, raising picture quality, and meets hardware and process in real time, the time delay one two field picture time, and does not affect video playback.

Accompanying drawing explanation

Fig. 1: the video defogging method capable principle schematic that the present invention is based on FPGA;

Fig. 2: the video mist elimination system composition schematic diagram that the present invention is based on FPGA;

Fig. 3: described video mist elimination system mist elimination degree regulates schematic diagram.

Embodiment

In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.

The present invention proposes to carry out mist elimination by FPGA to video.As shown in Figure 1, the method specifically comprises the steps:

Step S1: receive and store current frame video image M (x), simultaneously, detect the maximum in R, G, B value of each pixel in current frame video image and minimum value, obtain maximum matrix M ax (x) and minimum value matrix M in (x).R, G, B value of such as a certain pixel is respectively 223,114,51, then the maximum of this pixel is 223, and minimum value is 51.The maximum of each pixel forms maximum matrix M ax (x), and the minimum value of each pixel forms minimum value matrix M in (x).

Step S2: carry out mean filter to minimum value matrix M in (x), exports the mean filter matrix M of minimum value _ave(x).

Step S3: the mean value m of each minimum value in calculated minimum matrix _av.

Step S4: according to mean filter matrix and mean value calculation atmospheric transmissivity matrix L ₀.The formula calculating atmospheric transmission rate matrix is: L ₀=min (min (pm _av, 1.0) and M _ave(x), M (x)), M _ave(x)=average _sa(Min (x)).FPGA also can connect PC, and PC regulates FPGA by serial ports, to regulate mist elimination intensity p, thus adjustment atmospheric transmissivity matrix L ₀, realize the control to mist elimination intensity.

Step S5: calculate atmosphere light composition A according to maximum matrix, mean filter matrix and current frame video image.Atmosphere light composition computing formula is: $A=\frac{1}{2}(\max \left(\mathrm{Max}\left(x\right)\right)+\max \mathrm{Mave}\left(x\right))\left[\begin{array}{c}1\\ 1\\ 1\end{array}\right].$

Step S6: when receiving next frame video image, reads the RGB data of current frame video image, and is converted into yuv space data.

Step S7: calculate the Y value in the yuv space data after mist elimination according to the Y value in atmospheric transmission rate matrix and atmosphere light composition and yuv space data.Computational methods are: wherein Y' is the Y value in yuv space data after mist elimination, and Y is the Y value in the yuv space data before mist elimination.

Step S8: the yuv space data after utilizing the Y value in the yuv space data after mist elimination and the U value in yuv space data and V value to form mist elimination.

Step S9: the yuv space data after mist elimination are converted to RGB data, obtains the RGB data after current frame video image mist elimination.

When implementing of the present invention, the RGB data of current frame video image can be stored in outside DDR memory, the result of calculation of current frame video image can be applicable in the process to current frame video image, the violent saltus step of mist elimination is there will not be in the obvious video of image saltus step, more advanced than the method that the result of calculation of current frame video image is applied in the process of next frame video image.

Based on above-mentioned video defogging method capable, present invention also offers a kind of video mist elimination system based on FPGA, by FPGA101, mist elimination is carried out to current frame video image.Composition graphs 2 and Fig. 3, this system comprises: matrix builds module 2, mean filter module 3, mean value calculation module 4, atmospheric transmissivity matrix computations module 5, atmosphere light composition computing module 6, yuv space data conversion module 7, yuv space data construct module 9 and RGB data modular converter 10 after Y value computing module 8, mist elimination after mist elimination.Wherein:

Matrix builds module 2 in order to receive and to store current frame video image 1, meanwhile, detects the maximum in R, G, B value of each pixel in current frame video image and minimum value, obtains maximum matrix and minimum value matrix.

Mean filter module 3, in order to carry out mean filter to minimum value matrix, exports the mean filter matrix of minimum value.

Mean value calculation module 4 is in order to the mean value of minimum value each in calculated minimum matrix.

Atmospheric transmissivity matrix computations module 5 is in order to according to mean filter matrix and mean value calculation atmospheric transmission rate matrix.

Atmosphere light composition computing module 6 is in order to calculate atmosphere light composition according to maximum matrix, mean filter matrix and current frame video image.

Yuv space data conversion module 7, in order to when receiving next frame video image, reading the RGB data of current frame video image, and being converted into yuv space data.

After mist elimination, Y value computing module 8 is in order to calculate the Y value in the yuv space data after mist elimination according to the Y value in atmospheric transmission rate matrix and atmosphere light composition and yuv space data.

After mist elimination, yuv space data construct module 9 forms the yuv space data after mist elimination in order to utilize the Y value in the yuv space data after mist elimination and the U value in yuv space data and V value.

RGB data modular converter 10, in order to the yuv space data after mist elimination are converted to RGB data, obtains the RGB data after current frame video image mist elimination.Current frame video image is stored in DDR memory, and this DDR memory can adopt DDR2RAM102.

The formula calculating atmospheric transmission rate matrix is: L ₀=min (min (pm _av, 1.0) and M _ave(x), M (x)); Wherein, p is, m _avfor the mean value of minimum value each in minimum value matrix; M _avex () is mean filter matrix; M (x) is current frame video image; M _ave(x)=average _sa(Min (x)), wherein, Min (x) is minimum value matrix.The computing formula of atmosphere light composition is: $A=\frac{1}{2}(\max \left(\mathrm{Max}\left(x\right)\right)+\max \mathrm{Mave}\left(x\right))\left[\begin{array}{c}1\\ 1\\ 1\end{array}\right].$

System also comprises PC 103, and PC 103 regulates FPGA by serial ports, to regulate the value of mist elimination intensity p, thus adjustment atmospheric transmissivity matrix L ₀, realize the control to mist elimination intensity.

Native system only needs storage one frame video, just can meet the function of fast in real time mist elimination, improves picture quality, and meets hardware and process in real time, a time delay one two field picture time, can not affect video playback.The present invention uses FPGA101 for process main body, can be applicable to multi-signal I/O mode, can be applied in different video source and display occasion, have easy installation, compatibility by force, to the less demanding feature of operating personnel.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. based on a video defogging method capable of FPGA, it is characterized in that, described method carries out mist elimination by FPGA to current frame video image, comprises the steps:

Receive and store current frame video image, meanwhile, detecting the maximum in R, G, B value of each pixel in described current frame video image and minimum value, obtaining maximum matrix and minimum value matrix;

Mean filter is carried out to described minimum value matrix, exports the mean filter matrix of minimum value;

Calculate the mean value of each minimum value in described minimum value matrix;

According to described mean filter matrix and described mean value calculation atmospheric transmission rate matrix;

Atmosphere light composition is calculated according to described maximum matrix, mean filter matrix and described current frame video image;

When receiving next frame video image, read the RGB data of described current frame video image, and be converted into yuv space data;

The Y value in the yuv space data after mist elimination is calculated according to the Y value in described atmospheric transmission rate matrix and atmosphere light composition and described yuv space data;

Yuv space data after utilizing the U value in the Y value in the yuv space data after mist elimination and described yuv space data and V value to form mist elimination;

Yuv space data after described mist elimination are converted to RGB data, obtain the RGB data after described current frame video image mist elimination.

2., as claimed in claim 1 based on the video defogging method capable of FPGA, it is characterized in that, described current frame video image is stored in DDR memory.

3. as claimed in claim 1 based on the video defogging method capable of FPGA, it is characterized in that, the formula calculating atmospheric transmission rate matrix is: L ₀=min (min (pm _av, 1.0) and M _ave(x), M (x)); Wherein, p is mist elimination intensity, m _avfor the mean value of each minimum value in described minimum value matrix; M _avex () is described mean filter matrix; M (x) is described current frame video image; M _ave(x)=average _sa(Min (x)), wherein, described Min (x) is described minimum value matrix.

4., as claimed in claim 1 based on the video defogging method capable of FPGA, it is characterized in that, also comprise the steps:

PC regulates described FPGA by serial ports, to regulate the value of described p.

5., as claimed in claim 3 based on the video defogging method capable of FPGA, it is characterized in that, the computing formula of described atmosphere light composition is: $A=\frac{1}{2}(\max \left(\mathrm{Max}\left(x\right)\right)+\max \mathrm{Mave}\left(x\right))\left[\begin{array}{c}1\\ 1\\ 1\end{array}\right].$

6., based on a video mist elimination system of FPGA, it is characterized in that, described system carries out mist elimination by FPGA to current frame video image, comprising:

Matrix builds module, in order to receive and to store current frame video image, meanwhile, detects the maximum in R, G, B value of each pixel in described current frame video image and minimum value, obtains maximum matrix and minimum value matrix;

Mean filter module, in order to carry out mean filter to described minimum value matrix, exports the mean filter matrix of minimum value;

Mean value calculation module, in order to calculate the mean value of each minimum value in described minimum value matrix;

Atmospheric transmissivity matrix computations module, in order to according to described mean filter matrix and described mean value calculation atmospheric transmission rate matrix;

Atmosphere light composition computing module, in order to calculate atmosphere light composition according to described maximum matrix, mean filter matrix and described current frame video image;

Yuv space data conversion module, in order to when receiving next frame video image, reading the RGB data of described current frame video image, and being converted into yuv space data;

Y value computing module after mist elimination, in order to calculate the Y value in the yuv space data after mist elimination according to the Y value in described atmospheric transmission rate matrix and atmosphere light composition and described yuv space data;

Yuv space data construct module after mist elimination, forms the yuv space data after mist elimination in order to utilize the U value in the Y value in the yuv space data after mist elimination and described yuv space data and V value;

RGB data modular converter, in order to the yuv space data after described mist elimination are converted to RGB data, obtains the RGB data after described current frame video image mist elimination.

7., as claimed in claim 6 based on the video mist elimination system of FPGA, it is characterized in that, described current frame video image is stored in DDR memory.

8. as claimed in claim 6 based on the video mist elimination system of FPGA, it is characterized in that, the formula calculating atmospheric transmission rate matrix is: L ₀=min (min (pm _av, 1.0) and M _ave(x), M (x)); Wherein, p is mist elimination intensity, m _avfor the mean value of each minimum value in described minimum value matrix; M _avex () is described mean filter matrix; M (x) is described current frame video image; M _ave(x)=average _sa(Min (x)), wherein, described Min (x) is described minimum value matrix.

9., as claimed in claim 6 based on the video mist elimination system of FPGA, it is characterized in that, also comprise PC, described PC regulates described FPGA by serial ports, to regulate the value of described p.

10., as claimed in claim 8 based on the video mist elimination system of FPGA, it is characterized in that, the computing formula of described atmosphere light composition is: $A=\frac{1}{2}(\max \left(\mathrm{Max}\left(x\right)\right)+\max \mathrm{Mave}\left(x\right))\left[\begin{array}{c}1\\ 1\\ 1\end{array}\right].$