Research On The Detection Of Complex Parts And Related Key Technologies Based On Machine Vision

Due to low detecting accuracy, low detection efficiency and high cost factors, the traditional detection method for complex parts can't meet the needs of the modern production. Based on this, a non-contact measurement method based on machine vision is proposed in this paper. Taking diesel engine crankshaft of complex parts as test object and using the principle of visual inspection, the paper carries on the deep research on the degree of curvature of diesel engine crankshaft axis.First of all, according to the requirements of detection task and detection accuracy, the overall structure design of the visual detection system has been completed. Secondly, carry out calibration for the visual inspection system. Considering the noises produced inevitably in gathering the image, the image has been under the filtering and binarization processing. And based on the single structure characteristics of the diesel engine crankshaft image, a edge detection algorithm based on gray value difference has been put forward. Upon examination, compared with the traditional edge detection algorithm, the algorithm is precision and faster in detecting edge. Then, get the sub pixel edge image through the bilinear interpolation to improve the detection precision of the system. And by using the least squares, several pixel column section of crankshaft has been for the round curve fitting to get the center coordinates of each section. Finally, reuse the least squares to fit the line curve to calculate the straightness size and have a judgment on the bending condition of diesel engine crankshaft axis. Experiments show that the system is not only high precision, but also very fast.