Research And Application Of Image Defogging And Enhancement Technology

In haze or foggy weather,the image acquired outdoors has low definition,which makes the processing performance of the image processing system of the traffic department decrease,and the ability to recognize objects drastically reduces.Therefore,how to effectively defog and enhance the foggy image is a very practical research topic.This paper presents a fast defogging algorithm based on a single image.Based on the atmospheric light model used by predecessors,an improved filtering algorithm is introduced.After using atmospheric light model to test ambient light,we use bilateral filter with edge protection characteristics to smooth the dust concentration map,which can well preserve the edge of scene depth mutation.Secondly,the high-frequency noise can be effectively suppressed by using the partial median filter.We use the partial median filter to smooth the edge corners,so that we can avoid retaining a certain amount of fog around small targets,and retain the abrupt edges of building corners,so that the image defogging results are more robust.The experimental results show that the proposed algorithm can effectively avoid the generation of image oversaturation and halo effect.Defogging of images will inevitably result in image feature weakening.In some special scenarios,we also take images with weakened features.Image enhancement for feature-weakened images is the basis of many other image operations.A new image enhancement algorithm based on improved Lab color space Retinex algorithm is proposed in this paper.Based on the non-interference between the brightness and chroma components in Lab color space,we transform the simple pre-processed image into Lab color space.In this space,we use the improved Retinex algorithm to operate the separated brightness components.The improved algorithm proposed in this paper has better image enhancement effect and can solve the problem of halo artifacts.We apply the improved image de-fogging and enhancement algorithm to the license plate recognition system in foggy weather.Firstly,the improved image defogging and enhancement algorithm is used to preprocess the foggy license plate image.Then,the enhanced image is edge-detected to find the location of the license plate.Finally,the obtained license plate is segmented and recognized.The experimental results show that our improved image de-fogging and enhancement algorithm improves the accuracy of license plate recognition in foggy weather.