Calibration on Camera's Intrinsic Parameters Based on Orthogonal Learning Neural Network and Vanishing Points

作者：Ge, Dong-Yuan; Yao, Xi-Fan^*; Xiang, Wen-Jiang; Liu, En-Chen; Wu, Yong-Zuo

来源：IEEE SENSORS JOURNAL, 2020, 20(20): 11856-11863.

DOI：10.1109/JSEN.2019.2957134

• Cameras
• Calibration
• Neural networks
• Mathematical model
• Optical sensors
• Germanium
• Machine vision
• Vanishing points
• calibration
• intrinsic parameters
• orthogonal learning neural network

According to the orthogonal characteristics of the rotation matrix, a novel technique on camera's intrinsic parameters calibration is proposed in this paper. Firstly, two sets parallel lines in the space are designed which are perpendicular, and a corresponding equation set is obtained according to their coordinates in space and in image plane. Then an orthogonal learning neural network is designed, trained, and 8 elements of camera project matrix are obtained according to the eigenvector corresponding to the minimal eigen-value. Similarly, according to the solution requirement, another orthogonal learning neural network is designed, so the camera's principal coordinates (u0, v0) and the amplification factors kx and ky are obtained respectively. Finally the simulation results show that the proposed method has good robustness, and orthogonality test shows that the proposed method is of high precision.

• 单位
  广西大学; 邵阳学院