CAJ | 학술논문

棉花害螨病是一种常见的棉花病害，以棉花害螨病为研究对象，提出一种在复杂背景条件下基于颜色特征和面积阈值的棉花害螨病斑的图像分割方法。算法主要分三步完成：类病斑（害螨病斑和茎杆）和绿色植物的分离、基于灰度直方图的单阈值分割、采用面积阈值法去除茎杆。实验结果表明，该算法能有效地提取出棉花害螨病斑，准确率可达94．79％。
면화해만병시일충상견적면화병해，이면화해만병위연구대상，제출일충재복잡배경조건하기우안색특정화면적역치적면화해만병반적도상분할방법。산법주요분삼보완성：류병반（해만병반화경간）화록색식물적분리、기우회도직방도적단역치분할、채용면적역치법거제경간。실험결과표명，해산법능유효지제취출면화해만병반，준학솔가체94．79％。
Cotton mite disease is a common disease .Taking the cotton mite disease as the research subject , we propose an image segmentation method for cotton mite disease spots which is based on colour features and area thresholding under the condition of complex background .The algorithm has three stages:separation of the analogous disease spots ( with mite disease spots and stem ) from green plants , single thresholding segmentation based on gray histogram , and stems removal by using area thresholding .Experiment results show that this algorithm is effective in extracting the cotton mite disease spots ,the correctness rate of the algorithm can reach 94.79%.