Optimal voltage compensation method for focal plane avalanche detector based on improved random forest

Abstract

Abstract Working distance and background radiation greatly affect the signal-to-noise ratio of avalanche photodiode (APD) in the lidar detection system. The traditional method cannot adapt to a complex environment by offline compensation or pre-compensation according to the influence factors of the external environment. In this paper, an avalanche photodiode voltage compensation method based on the improved random forest is designed. Firstly, the distance image data is de-noised. Then the weight of each decision tree in the random forest was changed to improve the classification performance. The particle swarm optimization (PSO) algorithm was used to search for the optimal combination of parameters affecting classification accuracy and performance. Finally, the improved random forest algorithm is used to judge the current working state of APD at different distances, compensate for the bias voltage, and make APD work in the optimal state. The proposed method is compared with the k-nearest neighbor, support vector machine, and other commonly used classification algorithms, and the results verify the effectiveness of the proposed method.