A real-time video defogging method
A real-time video and video image technology, applied in image enhancement, instrumentation, computing, etc., can solve the problems of image quality degradation, high computational complexity, and foggy images without dehazing enhancement, etc., and achieve high computing speedup ratio, The effect of low computational complexity and good computational effect
Active Publication Date: 2019-02-19
XIAN AVIATION COMPUTING TECH RES INST OF AVIATION IND CORP OF CHINA
View PDF8 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0002] With the development of computer technology and video image technology, the imaging quality of airborne video sources is greatly reduced in complex environments such as smog, rain, snow, and dust, which greatly affects the human eye's perception and judgment of the environment.
Traditional video defogging algorithms are difficult to meet the requirements of airborne real-time processing due to their high computational complexity
However, general simple image enhancement algorithms cannot meet the performance requirements of video dehazing.
In many scenarios, it does not play a role in defogging and enhancing the output of foggy images, but instead reduces the quality of video images
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0020] Taking a certain type of DSP embedded system as an example, the method is executed by multi-core programming, and a fixed-complexity calculation method is used to perform real-time defogging and enhancement of video images, including the following steps:
[0021] In the first step, the input foggy image is scanned point by block according to multi-core parallelism, and the minimum value of the three channels of red, green and blue is obtained and stored as the minimum value image.
[0022] The second step is to use the Van Herk minimum filtering algorithm to filter the local minimum value of the minimum value image. Due to the separable characteristics of the minimum value filter, after the image is filtered in the row direction, it is filtered again in the column direction to obtain the minimum area. The value is stored as the initial value of the dark channel.
[0023] The third step is to perform guided filtering on the input dark channel initial value image, and use...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a real time video defogging method. According to the method, the demand for defogging performance is satisfied; the calculation complexity is relatively low and the real time performance is relatively high. The method comprises: firstly, with respect to an input foggy video image, carrying out minimum value solution on red, green and blue three channels pixel by pixel, then carrying out parallelized rapid local minimum value filtering, thus obtaining a dark channel initial image; successively carrying out blocking according to rows and columns, filtering to obtain a final dark channel image; secondly, carrying out histogram statistics, setting a dark channel threshold value, dividing the input fogging image according to the dark channel threshold value, only counting the average colors of the points with corresponding dark channel numerical values more than the dark channel threshold value as atmosphere color parameters; and finally, carrying out reverse solution according to a dark channel physical model, thus obtaining the defogged enhanced video image as the output.
Description
Technical field: [0001] The invention belongs to the field of computer video processing, in particular to a method for real-time video defogging enhancement. Background technique: [0002] With the development of computer technology and video image technology, the imaging quality of airborne video sources is greatly reduced in complex environments such as smog, rain, snow, and sand, which greatly affects the perception and judgment of the environment by the human eye. Traditional video defogging algorithms are difficult to meet the requirements of airborne real-time processing due to their high computational complexity. However, general simple image enhancement algorithms cannot meet the performance requirements of video defogging. In many scenarios, it does not have the effect of defogging and enhancing the output of foggy images, but degrades the quality of video images. Therefore, there is a need for a video data processing method that not only meets the requirements of...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T5/00
CPCG06T5/003
Inventor 程岳文鹏程李亚晖张东红王海鹏白林亭
Owner XIAN AVIATION COMPUTING TECH RES INST OF AVIATION IND CORP OF CHINA