基于机器视觉的胶管尺寸测量系统
Machine Vision based Measurement System of Rubber Hose Size

在对胶管进行质量检验时，针对人工使用量具对胶管进行尺寸测量出现的效率低、易受主观影响等问题，本文设计了一种基于机器视觉的胶管尺寸测量系统，实现对胶管内径和壁厚的测量。进行内外圆检测时，使用改进的区域生长法，不同生长梯度方向采用不同顺序扫描；进行内径测量时，由于胶管存在挤压易变形的特殊性质，内径测量误差较大，本文选择测量其内圆面积来判断内径是否达到检验标准；进行壁厚测量时，应用同心圆性质、垂直平分线的性质及勾股定理等原理计算得到各处胶管壁厚。实验结果表明：内径测量误差小于0.07 mm，壁厚测量误差小于0.05 mm，系统响应时间约为2 s，满足工业测量要求，有较好的实用性与应用前景。
In the quality inspection of the hose, for the artificial use of measuring tools to measure the size of the hose appears low efficiency, subject to subjective influence and other problems, this paper designed a machine vision based on the size of the hose measurement system, to achieve the inner diameter and wall thickness of the hose measurement. When the inner and outer circles are detected, the improved region growing method is used, and different growth gradient directions are scanned in different order. When the inner diameter is measured, the inner diameter measurement error is large due to the special property that the rubber hose is easy to be deformed by extrusion. In this paper, the inner circle area is measured to determine whether the inner diameter reaches the inspection standard. When measuring the wall thickness, the concentric circle property, the vertical bisector property and the Pythagorean theorem are used to calculate the wall thickness of various rubber tubes. The experimental results show that the inner diameter measurement error is less than 0.07mm, the wall thickness measurement error is less than 0.05mm, the system response time is about 2s, which meets the requirements of industrial measurement, and has a good practicability and application prospect.