Research of image RE/CAM integration technology of wheel-nave based on neural network
Yin Xue-pu · Journal of Machine Design · 2011
The image preprocessing was carried out by using synthetic method.The image characteristics were extracted by applying SOFM network,and its extraction ability was superior to that of BP network.By adding the momentum term to the new two layers BP network,the effect of image compression by using this network was enhanced.The new RBF-BP network was proposed,which inherits the advantages of the RBF and BP networks while eliminating the disadvantages,both the quality and efficiency of data repair to the image were improved significantly.The image RE/CAM integration system was developed by using Matlab and VC++,the NC codes and tool paths could be generated quickly and automatically based on image information,the size and shape detection as well as surface defect recognition could be carried out according to the image of parts.