【摘要】 针对传统的作物中心线提取方法易受作物种植间隙和噪声的影响,利用Meanshift算法能有效去除作物种植间隙和细小噪声的影响和Hough变换抗干扰能力强的特点,提出基于Meanshift和Hough变换的秧苗行中心线提取方法。该算法首先对水稻图像进行超绿变换(2G-R-B)和阈值处理,成功分离了目标与背景、减少了噪声并且提取出秧苗行的二值图像;然后对二值图像应用Meanshift算法去除叶片间的间隙和细小的噪声;最后应用Hough变换提取出秧苗行中心线。实验结果表明:相对传统的作物中心线提取方法,该算法获得的结果鲁棒性好,精确度高。更多还原

【Abstract】 Traditional center line extraction method of crops is easily affected by the crop planting clearance and the noise.Meanshift algorithm can effectively remove influence of crop clearance and small noise and Hough transform has a strong anti-interference ability.Thus,this paper proposes center line extraction method of rice seedling based on Meanshift transform and Hough transform.This algorithm first carries out 2G-R-B transform and threshold processing for rice image,successfully separates the target and background,reduces noise and extracts binary image of rice seedling row.Then,Meanshift algorithm is applied for binary image to remove gap among leaves and small noise.Finally Hough transform is applied to extract center line of rice seedling row.The results show that this algorithm can gain good robustness and high accuracy compared with traditional crops.更多还原