基于最佳模板匹配结合几何特征的土石料图像的去噪算法

In the image of earth and stone, the color of stone, sand and soil is mixed. The size of the stone is different. The sand and soil cover the surface of the stone, which is difficult to identify. There is a large amount of sand noise adhesion target between the targets. Aiming at the problem of sand noise adhesion target, the current representative template matching algorithm and geometric feature realize the separation of the adhesion target. The method has strong applicability to the image of the target size of the shape rule. The disadvantage is that for complex earth image rock and sand, Soil can\'t denoise very well. Combined with the image feature information, this paper proposes the best template matching and geometric feature of the stone material denoising algorithm in the earth and stone image to quickly separate the ore, sand and soil information in the earth and stone. The experimental results show that compared with the template matching and geometric feature method, the method is accurate denoising, which greatly improves the ore recognition accuracy.