论文部分内容阅读
用综合法做图像预处理。用SOFM网络提取图像特征,其提取能力优于BP网络。对新型二层BP网络增添了动量项,增强了用该网络进行图像压缩的效果。提出了RBF-BP新型网络,它继承了RBF和BP两网络的优势且消除了劣势,对图像实施数据修补的质量和效率显著提高。用M atlab和VC++开发出图像RE/CAM集成系统,能依据图像信息快速自动生成NC代码和刀具轨迹,并可依据零件图像进行尺寸、形状检测和表面缺陷识别。
Using integrated approach to do image preprocessing. Using SOFM network to extract image features, its extraction ability is better than BP network. Added a momentum term to the new two-layer BP network, enhancing the effect of image compression using the network. This paper proposes a new RBF-BP network, which inherits the advantages of RBF and BP networks and eliminates the disadvantages. The quality and efficiency of image data repair are significantly improved. Using M atlab and VC ++ to develop the image RE / CAM integrated system, the NC code and the tool path can be quickly and automatically generated according to the image information, and the size, shape detection and surface defect identification can be performed according to the part image.