摘要
针对大区域田间复杂背景下植物病害远程识别中的叶片病斑检测难问题,提出一种基于改进Bernsen二值化算法的植物病害远程检测方法。通过物联网采集不同区域的植物叶片图像,根据在RGB和HIS颜色空间中叶片病斑与正常叶片和背景的色调差异的特点,利用改进Bernsen二值化算法分别在图像的R、G、B、H 4个颜色通道上提取病斑,然后进行病斑图像融合,得到病斑图像。采用该方法对多幅物联网视频植物病害叶片图像进行病斑分割。实验结果表明,该算法在复杂背景环境下能够有效分割植物病斑图像,去除大量复杂背景,得到病斑图像。该方法能够为大区域植物病害远程智能监控系统提供技术指导。
As for the diffcultity of leaf spot disease detection in plant remote identification under complex background of large field,a remote detection method of plant disease was proposed based on improved Bernsen binary algorithm. The disease leaf images were collected by IOT from different areas. According to the different characteristics of color subspace of RGB and HIS of disease leaf and normal leaf and background colors,the spot images were extracted by the improved Bernsen binary algorithm from the four color channels of R,G,B and H,respectively. Then the spot images were obtained by spot image fusion. The proposed method was applied to segment several plant disease leaf images of agricultural IOT. Results showed that the improved algorithm could effectively segment the plant disease images in the complex background environment,remove a large complex background,and obtain the spot image. The proposed method can provide technical guidance for the remote intelligent monitoring system of plant disease in large areas.
出处
《广东农业科学》
CAS
2016年第12期129-133,F0002,共6页
Guangdong Agricultural Sciences
基金
国家自然科学基金(61473237)
陕西省自然科学基础研究计划项目(2014JM2-6096)
