论文部分内容阅读
为了对某全景式航空相机所产生的前向像移进行补偿,对前向像移产生的原理进行分析,并对像移补偿的方法进行研究。根据该型相机的工作原理,找到前向像移产生的机理。用神经网络模型(NNM)对系统时域响应的非线性特性进行辨识。根据NNM给出的扫描反射镜补偿角速度的预测值,通过最优算法调整控制信号,使被控对象跟踪理想补偿角速度的性能达到最优。该方法在该型相机前向像移补偿上克服了传统PID控制方法中所存在的不足,即精度低、响应慢、适应性差。实验结果表明:采用神经网络预测控制,扫描反射镜的实际补偿角速度输出跟踪理想补偿角速度的性能良好,优于传统方法。该方法可以应用于该型航空相机的像移补偿中。
In order to compensate for the forward image shift produced by a panoramic aerial camera, the principle of forward image shift is analyzed and the method of image shift compensation is studied. According to the working principle of this type of camera, find the mechanism of forward image shift. The neural network model (NNM) is used to identify the nonlinear characteristics of the system time-domain response. According to the prediction value of compensating angular velocity of scanning mirror given by NNM, the control signal is adjusted by the optimal algorithm so as to optimize the performance of the controlled object to track the ideal compensated angular velocity. This method overcomes the shortcomings of the traditional PID control methods in the forward image shift compensation of this type of camera, namely, low precision, slow response and poor adaptability. Experimental results show that the neural network predictive control and scanning mirror’s actual compensated angular velocity output can track the ideal compensated angular velocity well, which is better than the traditional method. This method can be applied to image shift compensation of this type of aerial camera.