基于改进Canny的水稻倒伏区域提取方法  

Rice Lodging Region Extraction Method Based on Improved Canny

作　　者：柴俊[1] 王浩陈 CHAI Jun;WANG Haochen(Wuxi Machinery and Electron Branch of Jiangsu Union Technical Institute,Wuxi 214028,China;School of Automation,Jiangsu University of Science and Technology,Zhenjiang 212100,China)

机构地区：[1]江苏联合职业技术学院无锡机电分院,江苏无锡214028 [2]江苏科技大学自动化学院,江苏镇江212100

出　　处：《现代信息科技》2025年第2期12-15,共4页Modern Information Technology

基　　金：江苏省省重点研发计划产业前瞻性与共性关键技术重点项目(BE2018007)。

摘　　要：为衡量农作物受灾后倒伏带来的产量与经济损失,利用改进的图像边缘检测技术来获取倒伏区域面积。在现有Canny算子边缘检测的基础上,采用中值滤波替代高斯函数来平滑处理图像。再采用自适应双阈值和广义链区分真伪边缘点、消除伪边缘点,最后通过线性拟合和霍夫变换获得边缘直线。由于倒伏区域和正常区域存在明显的分界现象和倒伏纹理增多的特征可准确提取倒伏区域。试验结果表明,五个实验地块倒伏面积提取误差只有9.2%,9.4%,8.9%,7.9%和8.7%。In order to measure the yield and economic losses caused by crop lodging after disasters,an improved image edge detection technology is used to obtain the area of the lodging area.Based on the existing Canny operator edge detection,the median filtering is adopted to replace the Gaussian function for smoothing the image.Then,the adaptive dual threshold and the generalized chain are used to distinguish between true and false edge points and eliminate the false edge points.Finally,the edge straight lines are obtained through linear fitting and the Hough transform.Due to the obvious boundary phenomenon between the lodging area and the normal area and the increased features of the lodging texture,the lodging area can be accurately extracted.The experimental results show that the extraction errors of the lodging area in the five experimental plots are only 9.2%,9.4%,8.9%,7.9%,and 8.7%.

关 键 词：倒伏作物 图像识别 边缘检测 改进Canny 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]