【摘要】 为减少揭膜机揭膜过程中对地垄边膜的损害，对揭膜机的土膜识别系统进行了设计。首先，对电动揭膜机的整体结构和工作原理进行分析；然后，对揭膜机的摄像头进行畸变校正，并对田间土膜图像的多种处理方式进行对比分析，确定最适合本系统的图像分割方式为K-means++法；最后，进行图像降噪，识别并提取地膜轮廓信息。试验测得图像的特征值范围为面积4500～20000mm~2,面积占比4%～22%,周长250～950mm, 3个条件同时满足即为土膜图像。在土壤板结状态下进行试验，结果表明：系统的平均识膜率大于95%,地膜的损伤率在5%以下，边膜上的土壤清除率到达82%。更多还原

【Abstract】 In order to reduce the damage of ridge edge film in the process of film uncovering, the soil film recognition system of film uncovering machine was designed. Firstly, the overall structure and working principle of the electric film unloader are analyzed, and then the distortion correction of the camera of the film unloader is conducted. After comparative analysis of various processing methods of field soil film image, k-means ++ method is determined as the most suitable image segmentation method for this system. Meanwhile, image noise reduction is carried out to identify and extract the contour information of mulch film. The range of eigenvalues of the experimentally measured image is area 4500～20000 mm~2, the area ratio is 4%～22%, the perimeter is 250～950 mm, it’s the soil film image. The test was carried out under the condition of soil compaction. The average film recognition rate of the system is more than 95%, while the damage rate of the film is less than 5%, and the soil clearance rate of the edge film reaches 82%.更多还原