融合图像处理的大粒度籽棉异纤清理系统设计

Design of cleaning system for large size seed cotton foreign fiber incorporating image processing

针对机采籽棉中滴灌带与地膜等大粒度异纤难以清理的难题,提出一种基于图像处理的籽棉异纤检测与清理系统,系统采用机械与风力技术设计了清除长杂及地膜残片的清理结构,同时设计了棉花异纤图像检测算法。采用高斯滤波对异纤图像平滑处理,通过一阶偏导的有限差分求取梯度的幅值与方向,并采用双阈值法对异纤中大粒度的残、碎片图像边缘进行分类识别与清除。实验结果表明,综合运用机械式清理与视觉检测清除的方法,能够快速准确检测并清理滴灌带与地膜等大粒度异纤,有效提高棉花加工企业异纤清除的效率与效果。

The system adopts mechanical and wind technologies to design the cleaning structure for removing long debris and mulch debris,and also designs the algorithm for cotton foreign fiber image detection,adopts Gaussian filter to smooth the foreign fiber image,calculates the amplitude and direction of gradient by the finite difference of first-order partial derivative,and the double-threshold method is used to classify and remove the residual and fragment image edges with large particle size in the foreign fiber.The experimental results show that the combined use of mechanical cleaning and visual inspection and removal method can quickly and accurately detect and clean the large-size foreign fibers such as drip irrigation tape and mulch,and effectively improve the efficiency and effectiveness of foreign fiber removal in cotton processing enterprises.