【摘要】 设计了一种基于机器视觉的皮棉异性纤维剔除系统,研究了异性纤维在线检测和剔除技术。为满足在线检测要求,提出了基于图像形态学理论的二进制图像腐蚀改进算法和高速气流喷吹的剔除方法。试验表明,在清洁皮棉的生产率为1000kg/h时,改进的图像处理算法对异性纤维达到95%的识别率,高速气流剔除异性纤维时,落棉率可以控制在4%。更多还原

【Abstract】 A new kind of foreign fiber separator was developed based on machine vision and the critical techniques such as on line trash detection and separation were analyzed. In order to meet the time limit of on line detection, a revised pattern recognition algorism derived from the binary morphological erosion theory was introduced. As a separation tactics, we used air spray to blow foreign fibers away from the cotton layer. The experiment based on these techniques showed that the precision of foreign fiber detection could reach 95% and the gin cotton loss could be controlled within 4%.更多还原