基于亮度均衡化和仿射变换的交通标志图像修复研究OA

Autonomous driving technology is the current research hotspot of artificial intelligence,in which the correct recognition of traffic signs is an important judgment basis for machines,but traffic sign images in natural scenes have a series of problems such as color deviation,shape deformation and occlusion,resulting in inaccurate recognition.For this problem,a method of brightness equalization and affine change is proposed,first using HSV color space,extracting the corresponding color,and then histogram equalization method for luminance information to solve the problem of color deviation.Subsequently,affine transformation is used to correct the shape of the triangle and circle images respectively,and to remove the background.The experimental part uses HOG+SVM detection method with low machine resource consumption for detection,and the results show that compared with traditional methods,this method has an improvement of about 20%in recognition rate.This proves that targeted restoration of traffic sign images has significant implications for improving recognition rates,as it consumes less machine resources and provides new ideas for current traffic sign recognition methods such as deep learning that consume more machine resources.