基于神经网络与遗传算法的图像轮廓序列匹配技术

切片轮廓图像匹配是实现基于层析图像的逆向工程产品反求的关键技术之一。该文针对经典算法应用于复杂产品样件切片轮廓图像引起的“难匹配”或“误匹配”问题,在建立物体轮廓森林的基础上,应用智能模式识别技术,将切片轮廓视为不同模式,设计一个三层的BP网络,对切片轮廓进行模式划分,并处理多分枝问题,实现轮廓匹配;应用遗传算法优化网络参数,提高算法效率。结合典型算例进行验证,取得了较好效果,该算法具有适应性强、准确性高的优点。

The Planar Contours Matching Technology Based on Neural Network and Genetic Algorithm

Slice contour matching is one of the key technologies in the reverse engineering based on slice image.Aimed at the hard or false matching problem when applying the classical algorithm to the slice contour image of complex products,this paper applies the intelligent pattern recognition technology to realize the matching and process the branching on the basis of the contour forest.In this algorithm,the contours are regarded as different patterns and a tri-layer BP network is designed to recognize the patterns.To improve the recognition efficiency,the genetic algorithm is also applied to optimize the parameters of the network.With the typical example,the algorithm is validated and it has the virtues of high fitness as well as high veracity.