中文摘要：

传统多传感器融合的刀具磨损检测方法，通过依据特征变换的特征降维方法，完成多传感器融合的刀具特征降维处理，其对特征的描述性差，检测效率低。因此，设计基于多视觉特征融合技术的高速机床刀具状态视诊系统，该系统通过固定摄像机A采集整体刀具图像，可控摄像机B采集刀头图像。两个摄像机的视频图像都输入到图像采集卡中的数据采集电路进行处理。系统通过数据采集电路获取刀具图像数据后，将数据传递给数据处理模块进行存储和模/数转换等处理。采用STC89C52单片机设计显示报警模块，用于显示刀具磨损状态。系统实现部分给出了系统软件流程图，并通过BP神经网络方法融合多视觉特征信息，检测高速机床刀具的磨损情况。实验结果表明，所设计系统可准确检测出刀具的磨损状态，具有较高的检测精度和鲁棒性。

英文摘要：

The cutter wear detection method of the traditional multi?sensor fusion is used to perform the cutter feature dimension reduction according to the feature dimension reduction method of the feature transform,which has poor feature description and low detection efficiency. Therefore,a high?speed machine cutter condition diagnosis system based on multi-vision feature fusion technology was designed,in which the images of the whole cutter are collected by the fixed camera A,and the images of the cutter bit are collected by the controllable camera B. The video images of the two cameras are input into the data acquisition circuit in the image acquisition card for processing. After acquiring the cutter image data through the system′s data acquisition circuit,the data is transmitted to the data processing module for storage,A/D conversion,etc. The single chip STC89C52 is adopted to design the display alarm module to display the cutter wear condition. The system software flow chart is given in System Implementation paragraph in this paper. The multi-vision feature information is fused with the BP neural network method todetect the wear conditions of the high-speed machine cutter. The experimental results indicate that the system can detect the cutter wear condition accurately,and has high detection accuracy and robustness.