Image defogging method and device

Abstract

The invention discloses an image defogging method and device, which includes a dark channel calculation module, a histogram statistics module, an ambient light calculation module, a transmittance calculation module, a storage module, a transmittance selection module, a filtering module and a defogging processing module; The image is divided into multiple blocks, the dark channel value of each block in the RGB component is calculated, the histogram is then calculated, and the ambient light value is calculated based on the histogram. And calculate the transmittance of the small, medium and large areas. After filtering, use the dark channel prior algorithm model to perform defogging processing to obtain the dehazed image data. Based on the existing dark channel prior algorithm model, the present invention optimizes parameters such as transmittance and ambient light, so that the block effect and edge effect of the image are better solved, and the restoration process is The dehazed image has high definition.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image defogging method and device. Background technique [0002] With the development of digital image processing technology and VLSI, high resolution and high processing speed have become the development trend of video image technology. High-definition and high-quality images play a vital role in traffic monitoring, information science, and military fields. However, in fields such as outdoor traffic monitoring, the camera system is often affected by bad weather, and weather such as fog, haze, rain and snow directly reduces the clarity of the image. Therefore, it is particularly important to perform dehazing processing on video images in hazy weather. [0003] In the prior art, the processing of foggy images mainly focuses on the algorithm level of software development on the PC platform. Although these algorithms can achieve the purpose of defogging, becau...

AI Technical Summary

Problems solved by technology

Although these algorithms can achieve the purpose of defogging, because they are all software processing, they take a long time and cannot meet the high-speed and high-quality requirements of real-time monitoring chips such as ISP ((Image Signal Processing, Image Signal Processing)

Although some scholars have tried to realize the dehazing algorithm in FPGA (Field-Programmable Gate Array, Field Programmable Gate Array) and other hardware, they have degraded their own dehazing algorithm. Fog effects are bound to be affected

[0004] The defogging algorithm based on dark channel prior is recognized as the most effective and general algorithm, but most of the defogging methods and devices based on this algorithm model also have the problem of image block effect and edge effect, making The image transition is not smooth, and the quality of the restored defogged image cannot meet high requirements

Images