摘要
根据二维光电位置敏感器件 (PSD)的工作原理 ,分析了影响PSD线性度的主要因素 ,提出了一种用神经网络对PSD进行非线性修正的方法。以PSD的输入输出数据对作为样本训练的神经网络 ,利用神经网络所具有地能够以任意精度逼近非线性函数的能力 ,实现PSD的输出与实际光点位置之间的映射 ,在神经网络的输出端得到线性响应。该方法的优点是不需要很大的数据存储量即可得到很好的修正效果。结果表明 。
The main causes of non linearity of PSD are analyzed according to the operation principle of two dimension photoelectric position sensitive detector (PSD). A non linear correction method for the PSD using artificial neural networks (ANN) is presented. The ANN is used as a corrector that is trained by learning algorithm through the input/output data pairs of PSD. Utilizing feature of its powerful ability for function approximation, the ANN can perform the mapping between the PSD's readings and the light spot actual positional information. This enables the output signal of ANN to provide a linear response. The main advantage of this method is that an ideal effect can be realized by using only fewer memory requirements. The result shows that the PSD with a proper ANN correction can provide a high linearity over a wide position range.
出处
《光学技术》
CAS
CSCD
2002年第2期174-175,178,共3页
Optical Technique
基金
浙江省教育厅基金资助项目
