Research On Low Illumination Image Detection Based On Image Enhancement Algorithm

Low-light image enhancement algorithm is an important technology to improve image quality.At present,low-light image enhancement technology has achieved remarkable results in enhancing image brightness,but there are still several problems: 1)After the nonuniform illuminance image is processed by a low illumination image enhancement algorithm,some areas can easily expose or undergo exposure problems.2)After the low image is processed by a low illuminance image enhancement algorithm,the noise of the image itself is easily enhanced.3)After the low image is processed by a low illuminance image enhancement algorithm,the image is easy to show color distortion.These problems make the characteristic information of the target of the real scene,which is designed for these issues,and the specific research content is as follows:1.A multi-threshold block low-illumination image enhancement algorithm fused with genetic algorithm is designed to solve the problems of uneven illumination and blurred details of low-illumination images.First,the optimal segmentation threshold of the luminance feature map of the input image is obtained by genetic algorithm,and the image is divided into multiple sub-images with different exposure levels according to the obtained threshold,and the brightness of each sub-image is improved according to the complexity of each sub-image;In order to enhance the texture information and contour information of the image,an anti-sharpening mask algorithm and a guide filtering algorithm are introduced,and a multi-scale fusion algorithm is introduced in order to fuse the enhanced image texture information and the image profile information and the brightness.The brightness of the low illumination image treated with this algorithm is reduced,and the detailed information in which the underlying area under the image is not exposed.The AG value of the enhanced image increased by 1.69,and the EME value increased by 19.28,the EN value increased by0.69.2.A low-light image enhancement algorithm based on human vision is designed to solve the problems of color distortion and obvious noise in low-light image enhancement results.First,the information of the H and S channels of the image HSV space is converted into a weight map to adjust the brightness information of the V channel of the image,so that the color of the image is more in line with the visual characteristics of the human eye.Further,in order to filter the noise of the image,the algorithm introduces the guiding filter algorithm,and finally obtains the output image with normal brightness and vivid color.The AG value of the enhanced image increased by 2.45,and the EME value increased by 13.20,the EN value increased by 0.22.3.In order to further verify the application value of the enhancement algorithm,the two enhancement algorithms designed in this paper are combined with defect detection algorithm and text detection algorithm,and a set of tunnel safety inspection scheme suitable for harsh lighting environment is designed.By comparing the detection results of tunnel defect targets and text information in the low illumination images before and after enhancement by the tunnel safety inspection software,it is found that using the enhancement algorithm proposed in this paper to preprocess the images before detection can improve the detection ability of the model,and then verify the feasibility of the tunnel safety inspection scheme in this paper.