Visual Detection for Non-Ferrous Metal Ingots With Wavelet Denoising and Contour Corner Extraction | IEEE Conference Publication | IEEE Xplore

Visual Detection for Non-Ferrous Metal Ingots With Wavelet Denoising and Contour Corner Extraction


Abstract:

To realize the intelligent grinding of non-ferrous metal ingots, visual detection algorithms are required to provide the information of metal ingots for robotic manipulat...Show More

Abstract:

To realize the intelligent grinding of non-ferrous metal ingots, visual detection algorithms are required to provide the information of metal ingots for robotic manipulators. However, metal ingots in moving molds on ingot casting lines are accompanied by complicated backgrounds and continuous changes, which demands higher accuracy and stronger robustness of visual detection. Based on two kinds of metal ingots and the analysis of metal ingot images, a wavelet-based method with an improved threshold and threshold function is utilized to reduce noises. Afterward, the edge information of metal ingots is obtained by edge line segment detection, and the detected edges are selected according to their positions and lengths. Moreover, in order to extract the contour corners of metal ingots, basic contours are detected and further fitted into metal ingot contours. Finally, experimental results verify that the proposed metal ingot detection algorithm achieves satisfactory accuracy, robustness, and real-time performance for different images and metal ingots.
Date of Conference: 01-03 July 2022
Date Added to IEEE Xplore: 18 August 2022
ISBN Information:
Conference Location: Tianjin, China

Funding Agency:


References

References is not available for this document.