Generation of grinding area for vehicle surface maintenance based on convex hull algorithm
-
摘要: 针对传统汽车漆面维修以人工为主，存在劳动强度高、效率低、一致性差等问题，结合人工打磨经验，在提取受损车面轮廓的基础上，设计生成受损车面维修打磨区域算法. 首先，综合运用边缘检测算子、曲线近似算法对受损漆面进行轮廓提取. 然后，运用凸包算法预生成轮廓，再对轮廓改进生成最终车面维修打磨区域. 通过比较人工和算法生成后的打磨区域可知，算法生成的打磨区域精度较高，可满足基本打磨要求，该算法为实现自动化打磨提供了理论依据.Abstract: In a traditional industry, automotive paint repair is dominated by manual labor, which has problems such as high labor intensity, low efficiency and poor consistency. Combining the experience of manual sanding, the algorithm of generating the damaged car surface repair sanding area was designed based on the extraction of the damaged car surface contour. Firstly, the edge detection operator and convex hull algorithm were used to extract the contour of the damaged paint surface, then the convex package algorithm was used to pre-generate the contour, and then the contour was improved to generate the final car surface repair sanding area. By comparing the actual manual sanding area and the algorithm-generated sanding area, it shows that the sanding area generated by the algorithm is more accurate and meets the basic sanding requirements, which can provide a theoretical basis for the realization of automated sanding.
-
Key words:
- automotive paint repair /
- convex hull algorithm /
- sanding area generation
-
图 14 车面维修打磨区域生成图对比
Figure 14. Generation maps comparison of car surface repair and grinding area
表 1 人工打磨图和算法图对比结果
Table 1. Comparison results of manual polishing and algorithm maps
% 图像 S原始图/S人工打磨图 S原始图/S算法图 
15.7 73.0 
28.3 86.0 
下载: 导出CSV
-
[1] NGENDANGENZWA B. Defect detection and classification on painted specular surfaces [D]. Umeå: Umeå University, 2018. [2] AKHTAR S, TANDIYA A, MOUSSA M, et al. An efficient automotive paint defect detection system[J] . Advances in Science, Technology and Engineering Systems Journal,2019,4(3):171 − 182. [3] ZHOU Q B, CHEN R W, HUANG B, et al. Deep Inspection: Deep leaning based hierarchical network for specular surface inspection[J] . Measurement,2020,160:107834. [4] 韩双旺, 崔兆顺, 鲍丽红, 等. 基于目标图像的提取与测量[J] . 上海工程技术大学学报,2007,21(4):345 − 350. doi: 10.3969/j.issn.1009-444X.2007.04.015 [5] 廖鸿飞, 李鸿燕, 徐才恩, 等. 基于方向梯度的边缘检测算法及实现[J] . 上海工程技术大学学报,2017,31(1):16 − 19. doi: 10.3969/j.issn.1009-444X.2017.01.004 [6] BOUATTANE O, ELMESBAHI J, RAMI A. A fast parallel algorithm for convex hull problem of multi-leveled images[J] . Journal of Intelligent and Robotic Systems,2002,33(3):285 − 299. [7] ASGHAR H J, LI S J, PIEPRZYK J, et al. Cryptanalysis of the convex hull click hum an identification protocol [C]//Proceedings of the 13th International Conference on Information Security. Boca Raton: Springer, 2010: 24−30. [8] 包振华. 凸集及凸包形态学运算的几个性质[J] . 科学技术与工程,2003(1):69 − 71. doi: 10.3969/j.issn.1671-1815.2003.01.023 [9] 陈尹刚. 基于数学形态学图像处理算法研究[J] . 信息通信,2019(12):67 − 68. -
下载:
点击查看大图
图(14) / 表(1)
计量
- 文章访问数: 41
- HTML全文浏览量: 14
- PDF下载量: 8
- 被引次数: 0
