Introduction

Abstract

Image processing technology mainly includes image transformation, image restoration, image compression, image segmentation, target detection and recognition, and image classification. Currently, image segmentation technology has been widely used in industrial detection, medical image analysis, remote sensing earth observation, and other fields. Image segmentation involves multidisciplinary knowledge, including imaging principles, color theory, cognitive science, and machine learning techniques, among which machine learning is the most important. With recent innovations in science and technology, autonomous driving has become a research trend and a hot spot in the artificial intelligence and automotive fields. Autonomous driving based on the perception system can promote the development of transportation systems and effectively broaden the application field of science and technology. In order to evaluate the advantages and disadvantages of image segmentation algorithms, researchers have developed many evaluation indexes. By imitating the human visual perception, these evaluation indexes mainly depend on error rate, segmentation contour quality, region quality, and other similar factors.