Journal of Xidian University ›› 2019, Vol. 46 ›› Issue (5): 75-83.doi: 10.19665/j.issn1001-2400.2019.05.011

Previous Articles     Next Articles

Method for the detection of the piston side defect based on external contour registration

WANG Hongyan1,2,3,ZHU Limin4,ZHANG Panjie1,2,3,LI Jinping1,2,3()   

  1. 1. School of Information Science and Engineering, University of Jinan, Jinan 250022, China
    2. Shandong Provincial Key Laboratory of Network Based Intelligent Computing, University of Jinan, Jinan 250022, China
    3. Shandong College and University Key Laboratory of Information Processing and Cognitive Computing in the 13th Five-Year Plan,University of Jinan, Jinan 250022, China
    4. Binzhou Bohai Piston Co., Ltd., Binzhou 256602, China
  • Received:2019-03-11 Online:2019-10-20 Published:2019-10-30
  • Contact: Jinping LI


In order to detect the defects on the side of the piston, we propose an effective method for detecting the piston side defect based on the registration of piston contour according to the high similarity between the images of two pistons of the same type and specification under the same illumination and angle. The method is divided into four steps: first, we establish a standard template dataset by taking multi-angle images of standard pistons of different types and specifications under standard illumination conditions; second, we use the Scale Invariant Feature Transform (SIFT) algorithm to register the template image and the current image according to the piston contour features so as to find the exactly corresponding region of the two images; third, the corresponding regions of the two images are traversed with sliding windows of the same size to calculate such features as mean, variance, vertical projection and horizontal projection; finally, we determine whether there is a defect in the current window by comparing the features of two corresponding windows. The results show that the method can effectively detect the piston surface defect and determine the position of the defect, and that the accuracy rate is 94.78%, and it has strong practicability.

Key words: defect detection, piston surface, feature extraction, image registration

CLC Number: 

  • TP391