Detection Technology Of Flexible Shaft Defect Based On X-ray Images

In non-electric control aircraft, the flexible shaft is used to control the engine main pump regulator,and control various the engine working conditions, so its quality directly affects the flight control and safety of the aircraft. The domestic R & D in the flexible shaft has a large gap between foreign countries, also with no matching defect detection equipment,so detection technology of the flexible shaft defect is of significance to the healthy development of national defense industry. The Defect of flexible shaft is the lack of internal ball, whose detection has acquisition of long image, complex structure, many balls and other difficulties,and this paper made research to those difficulties combining industrial DR and digital image processing technology. The main works of the paper are:(1) As to the difficulty of acquiring long DR image of flexible shaft, two methods are proposed: image splicing method and scan imaging method. The former is to let the car go by a certain step length to collect a number of overlapping images, and finally splice those images to a long image. While the latter is to let the car has been moving, and the host computer take out and storage several columns from each frame concurrently and finally generate a long image directly. Experiments show that, for image splicing method, the steps of it are complicated, and the overlap rate is high and its feasibility is poor. While the scan imaging method is simple, and the quality of acquired long image is high. Therefore, the scan imaging method is selected, and the long image is preprocessed, which lays a good foundation for defect detection.(2) For the defect detection of the long images, first of all, the general areas of the balls are segmented by the method of mathematical morphology combined with threshold segmentation, and the upper and lower layers are separated, which is favorable for the counting and abnormal analysis of the balls; And then calculate the center of the balls and the spaces between balls in every floor, and the mode of the spaces with a certain threshold is set as a criterion for determining whether the ball is normal; Then statistics of the abnormal balls interval is made,and the two common anomalies,"a pseudo" and "a missing",are automatic processed; Finally, statistical results of three kinds of ball,"normal","missing" and "other abnormal" are given. By testing two flexible shafts, the results show that the accuracy and recognition rate of the algorithm are high, which greatly reduces the manual workload.(3) X-ray detection software of flexible shaft has developed making the use of MFC and OpenCV development library, whose interface is Ribbon style that function demands are focused on,and is convenient and beautiful. The software can mark the "normal","missing"and "other abnormal" three types of balls on the long image using green, red and yellow color,and give the corresponding Word detection report, which is intuitive and convenient.