基于粗糙集的迭代阈值选择的玉米苗期背景分割算法研究

摘要

将粗糙集与阈值迭代算法相结合,提出了基于粗糙集的迭代阈值选择背景分割算法并将其运用在玉米苗期图像上.利用粗糙集的模糊性刻画目标与背景的边界,用阈值迭代法选取阈值进而实现背景分割.仿真试验结果表明:传统的阈值迭代算法的分割耗时为0.469s,并且分割后图像存在大量的信息丢失；基于粗糙集的迭代阈值选择背景分割算法的分割耗时为0.374 s,分割图像清晰,能够快速、有效、准确的实现背景分割,并满足农业智能生产中实时性、高效性的要求.%According to rough set and iterative threshold, the algorithm based on rough set for iterative threshold selection was proposed first and applied for the image of com seedling stage. Fuzzy characterization of rough set was applied for the boundary between background and object, and iterative threshold was used to select threshold to segment. The experiment show that the segmention time of the classical iterative threshold is 0.469 s, the improved algorithm can rapidly, accurately and effectively perform segmentation and the segmention time is 0.374 s. It can satisfy the request of real-time agricultural intelligence production.