The utilization of reserved holes on the chimney wall for the lifting and locking of the chimney demolition robot can significantly enhance the working efficiency of the robot. A method is proposed for identifying and locating reserved holes on the chimney wall based on deep learning object detection technology. Firstly, the YOLOv5 deep learning network model is used to detect the approximate area of the reserved holes in the image, obtaining the coarse positioning information of the reserved holes. Then, the Gaussian filtering method is applied to remove isolated points and noise in this area. Subsequently, based on the coarse positioning information of the reserved holes, the Canny operator is used to extract the edge information of the reserved holes. Finally, the least squares-based elliptical fitting method is employed to obtain the center coordinates of the reserved holes, achieving precise positioning of the reserved holes and effectively reducing the impact of factors such as blurred target edges, similar object interference, and circular distortion on the positioning accuracy of the reserved holes. The research results show that the mAP@0.5% of the YOLOv5s model reaches 99.5%, with a detection speed of 6.5 ms. In the simulation environment, the center positioning error is less than 0.8 pixels, which can effectively improve the reliability and efficiency of the lifting operation of the chimney demolition robot.
LUDongfang, LIHandong. Research on the algorithm of the center locationing of a circle based on OpenCV[J]. Microprocessors, 2023, 44(1): 22-25. (in Chinese)
ZHENGJin, YANJiajie, WANGQingxia, et al. Accurate wafer alignment method based on feature recognition[J]. Journal of Donghua University(Natural Science Edition), 2025, 51(2): 198-205. (in Chinese)
[9]
PANS, WANGS, XUJ, et al. Sub-pixel position estimation algorithm based on Gaussian fitting and sampling theorem interpolation for wafer alignment[J]. Applied Optics, 2021, 60(31): 9607-9618.
ZHOULanfeng, LILing, FANGHua, et al. Improvement and application of lunar crater detection algorithm based on Hough transform[J]. Computer Applications and Software, 2023, 40(5): 266-271. (in Chinese)
LIUTianci, LIUGuihua, HULi, et al. Laser center-line extraction method based on normal guidance[J]. Journal of Applied Optics, 2023, 44(1): 211-218. (in Chinese)
[14]
SHENY, ZHANGX, CHENGW, et al. Quasi-eccentricity error modeling and compensation in vision metrology[J]. Measurement Science and Technology, 2018, 29(4): 045006.
[15]
XUW, LIQ, FENGH J, et al. A novel star image thresholding method for effective segmentation and centroid statistics[J]. Optik, 2013, 124(20): 4673-4677.
WANGYin, JIANGZheng, LIUBin. SIFT fast image matching algorithm with local adaptive threshold[J]. Chinese Journal of Liquid Crystals and Displays, 2024, 39(2): 228-236. (in Chinese)
[18]
ARAGUESR, GONZÁLEZA, LÓPEZ-NICOLÁSG, et al. Distributed relative localization using the multidimensional weighted centroid[J]. IEEE Transactions on Control of Network Systems, 2020, 7(3): 1272-1282.
CHIXiaobo, ZHANGWeijie, JIAXinchun, et al. Lightweight method for detecting insulator missing based on improved YOLOv5s[J]. Journal of Test and Measurement Technology, 2024, 38(1): 19-26. (in Chinese)
[23]
LIS, ZHANGJ, LIUB, et al. An algorithm to extract the boundary and center of EUV solar image based on sobel operator and FLICM[J]. Photonics, 2022, 9(12): 889.
[24]
ZHANGC, LIT, LIJ. Detection of impurity rate of machine-picked cotton based on improved canny operator[J]. Electronics, 2022, 11(7): 974.
[25]
NOGUEIRAM L, GREISN P, SHAHR, et al. Machine learning classification of surface fracture in ultra-precision diamond turning using CSI intensity map images[J]. Journal of Manufacturing Systems, 2022, 64: 657-667.