Research on AGV visual perception dynamic exposure algorithm based on gray entropy threshold difference value
Junbao Zheng, Zihan Zhang · 2022 4th International Conference on Advances in Computer Technology, Information Science and Communications (CTISC) · 2022
For the narrow dynamic range camera commonly equipped on outdoor AGV vehicles, when the traditional exposure algorithm is used in the environment with variable light intensity, sometimes over-exposure and under-exposure will occur, which makes it difficult to obtain balanced exposure effect. In order to solve the problem of poor adaptability of traditional exposure algorithm to AGV narrow dynamic range CMOS sensor, an exposure algorithm based on gray entropy threshold difference value is proposed in this paper. Firstly, according to the distribution characteristics of key road information in the image obtained by AGV, regions of interest in the image are divided to exclude the interference of background information on subsequent processing. Then, the gray entropy threshold difference value is obtained by combining the OTSU threshold and the one-dimensional entropy value of the image, and it is used as a final control parameters to dynamically adjust the exposure time. This parameter has the advantages of gray threshold selection method and image information entropy method, it can feedback the current image quality more reasonably. Experimental results show that the algorithm can effectively adjust exposure under different illumination intensity. The algorithm can not only get better exposure effect, but also has better adaptability and stability in comparison with the conventional algorithm.