论文部分内容阅读
提出以激光束的横模结构特别是基模所占的比例作为激光光束质量的评价依据,并研究了测量方法.建立了基于Hopfield神经网络原理的非线性网络,使用CCD采集激光束的光斑图像作为网络的输入,通过计算该网络的能量函数,实行调整训练,获得网络的动力学稳定状态,此时网络中各阶横模的比例即为测量结果.实验采集了一束光的多幅光斑图像,经预处理后输入该网络,可获得模式结构数据,其中基模分量为69%.利用所得结果合成一幅光斑,与输入的原光斑图像的相对误差为3.53%.
It is proposed that the transverse mode structure of the laser beam, especially the proportion of the fundamental mode, be used as the evaluation basis of the laser beam quality and the measurement method is studied. A nonlinear network based on Hopfield neural network is established. As the input of the network, the energy function of the network is calculated and adjusted and trained to obtain the dynamic state of the network, in which case the ratio of the transverse modes in each network is the measurement result. After pre-processing and inputting into the network, the model structure data can be obtained, in which the fundamental mode component is 69%, and the relative error between the input image and the original image is 3.53%.