The Investigation For Key Technologies Of On-line Mearing For NC Processing

In the CNC machining process, due to low detection efficiency, accuracy, detection process laborious and other complicated reasons,the traditional manual testing has not met the requirements of production process for testing speed and accuracy.The higher demand was put forward to the modern manufacturing in CNC machining process,so they should be faster and more exactitude to the workpiece testing. The CNC machining CCD detection was widely researched by the scholars at homes and aboard and it would become the inevitable trend of detection methods of CNC machining for its high efficiency and high precision, reliable performance and intelligent features.This paper based on the background of the development of the CNC machining technology and CCD testing technology, the existing technology and development results of CCD detection technology were combined with,the framework and key technologies of CCD detection system were deeply studied.The major contributions list as follows:1. The level, developed results and trends of CCD detection technology of CNC machining at domestic and abroad were summaried;2. The principle and the advantages and disadvantages of computer vision detection technology for linear CCD projection and imaging were deeply analyzed and compared,the testing scheme was proposed and the difficulty and critical technologies of the system implement was analyzed;3. A theoretical model of the optical system was established, the selection and design of high intensity light source , the choice of lens and the calculate for optical parameters were accomplished,the aim of high quality imaging of the workpiece was achieved.4. The hierarchical processing of image was actualized through the use of the FPGA + DSP structure simultaneously the requirements of high-speed acquisition and real-time processing was met;the modular embedded acquisition system was designed besides TCDl209D drive circuit and other key signal conditioning circuit was gained.5. The algorithms of image filtering, image enhancement and edge detection were studied based on the topographic characteristics of selected measurement objects (step axis),image contour and parameters were obtained by Matlab 7.0. Finally, the application of the BP neural networks in image recognition was researched for the further study in the future.