基于神经网络和遗传算法的光塑性图像处理

Processing of Photoplastic Images Based on Neural Network and Genetic Algorithm

本文论述了神经网络、遗传算法结合起来在光塑性图像处理中的应用 ,以楔横轧工序的光塑性图像处理为例 ,三次投射时光塑性图像中各点都有对应的等差线条纹级数和等倾线角度 ,首先利用神经网络建立起图像参数与等效应变之间关系的数学模型 ,然后用遗传算法对图像参数进行优化 ,优化结果与实际情况吻合良好。由此证明神经网络、遗传算法完全可以较好地优化光塑性图像参数 。

Neural Network and Genetic Algorithm were combined together and applied to processing of photoplastic images. By taking the processing of photoplastic images of a cross wedge rolling as an example, each of the points in the photoplastic images which were got through casting for three times had the corresponding isoclinics and isochromatics. Firstly, the mathematics model between the parameters of photoplastic images and equivalent strain was set up with neural network. Then, the parameters of photoplastic images were optimized with the genetic algorithm. The result of optimization is consistent well with the actual instance. Therefore Neural Network and Genetic Algorithm can be used for optimizing the parameters of photoplastic images well and they can be applied to processing field of photoplastic images.