Research On Multi-Parameter On-Line Detection Of Cylinder Linder

With the development of the economy and automobile industry, the popularity of vehicles continuously increase in daily life. There is a rising trend on the demand of cylinder linder of which annual requirements has been up to ten million. Relying on the artificial detection on cylinder linder, results in the low efficiency and high error rate. The introduction of a flexible and multi-parameter on-line detection system for cylinder linder has been a imperative to automobile manufacturer. In recent years, the universal digital signal processor(DSP) has made considerable progress, and got more and more favor for its data processing ability that exceed PC and most of the embedded system in the industrial dection.In order to improve the processing precision and efficiency of engine cylinder liner on-line detection system, at first, this paper analyses the hardware structure of TMS320DM642, constructed the cylinder liner visual detection system of DM642-centred, studied the principle of major modules including DM642peripheral circuit, memory expansion circuit, video acquisition circuit and completed the corresponding hardware design, software development and integrated debugging of the system. Then, this paper further studied the cylinder linder visual detection method based on DM642, completed the camera calibration, the compilation, simulation, performance analysis of image edge detection, geometric feature extraction, and optimized the algorithm of geometric feature extraction according to the feature of cylinder linder. With the combination of the research results and detection technology of cylinder linder, It has certain practical significance for the development of the automobile industry in our country.