Intelligent calibration of dual liner CCD intersection testing system based on Backpropagation neural network

Abstract In order to improve the measurement accuracy and calibration efficiency of the dual liner Charge Couple Device (CCD) intersection measurement system, a calibration technique based on Backpropagation (BP) neural network is proposed. This calibration method does not consider the parameters of the linear CCD camera and treats the entire dual liner CCD intersection measurement system as a whole for calibration. In the testing of the dual liner CCD intersection measurement system, the pixel coordinates of the measured target in the pixel plane of the linear CCD camera will change with the actual position of the measured target. The position of the measured target is continuously changed multiple times to construct a training sample set. The BP neural network is trained using the constructed training samples, and the strong nonlinear mapping ability of the BP neural network is utilized to obtain the mapping relationship between the pixel coordinates of the measured target and the actual coordinates of the measured target in the dual liner CCD intersection measurement system. A dual liner CCD intersection measurement system was constructed, and calibration experiments were designed to compare and analyse the experimental data. The results show that the proposed method can improve the testing accuracy and stability of the dual liner CCD intersection measurement system, while also improving calibration efficiency.