WCSE

Research on Vehicle Detection Algorithm based on Haar-like and CNN

Abstract— Vehicle detection algorithms exists in various application scenarios, including driverless car and traffic control. We proposed a vehicle detection method based on Haar-like and Convolutional Neural Networks to address the problems of traditional Adaboost algorithm for Haar features has low precision when detecting vehicles. In this paper we proposed a method which through the Haar-like and integral maps to extract target features, calculate eigenvalues, then uses Convolutional Neural Networks to discriminate and remove false targets. Experimental results prove: our algorithm proposed in this paper can effectively improve the precision of detection at a good recognition speed.

Index Terms—Vehicle detection, Haar-like feature, CNN, Adaboost algorithm.

Jinlong Chen, Yiming Jiang, Minghao Yang
Guangxi Key Laboratory of Cryptography and Information Security , Guilin University of Electronic Technology, CHINA
Guangxi Key Laboratory of Trusted Software , Guilin University of Electronic Technology, CHINA

[Download]

Cite: Jinlong Chen, Yiming Jiang, Minghao Yang, " Research on Vehicle Detection Algorithm based on Haar-like and CNN " Proceedings of 2020 the 10th International Workshop on Computer Science and Engineering (WCSE 2020), pp. 125-131, Shanghai, China, 19-21 June, 2020.