利用人工神经网络来直接学习图像信息与三维信息之间的关系 ,不需确定摄像机具体的内部参数和外部参数 ,也不需知道有关模型或参数的先验知识。在双目视觉的情况下 ,两摄像机的位置关系不需具体求出 ,而是隐含在神经网络之中。实验结果表明神经网络方法的标定结果令人满意 ,并成功地用于机器人曲线跟踪的实验 The use of artificial neural networks to directly learn the relationship between image information and three-dimensional information without the need to determine the camera’s specific internal parameters and external parameters, and does not need to know the prior knowledge of the model or parameter. In the case of binocular vision, the positional relationship between the two cameras need not be specifically calculated, but implicit in the neural network. Experimental results show that the method of neural network calibration results are satisfactory, and successfully used in robot curve tracking experiments