Research on Pedestrian Detection Algorithm Based on Deep Learning

Abstract

Abstract Pedestrian detection is a kind of computer vision technology which is used to judge whether there is a pedestrian in a given image or video and estimate its position accurately. Pedestrian detection technology is widely used in intelligent security monitoring, driverless and robot fields. Pedestrian detection technology has nearly 30 years of development. Although it has greatly improved in accuracy and speed, there are problems such as lighting, pedestrian individual differences, occlusion and multi-scale pedestrian targets in the actual application scenarios, which lead to more missed and false detection in pedestrian detection. The algorithm is not robust enough, which makes the actual deployment difficult.