基于机器视觉的酿酒葡萄叶幕厚度估算方法研究  

Research on the Estimation Method of Canopy Thickness of Grapevine Based on Machine Vision

作　　者：马聪 陈学东 MA Cong;CHEN Xue-dong(Institute of Agricultural Economy and Information Technology,Ningxia Academy of Agriculture and Forestry Sciences,Yinchuan,Ningxia 750002)

机构地区：[1]宁夏农林科学院农业经济与信息技术研究所,宁夏银川750002

出　　处：《安徽农业科学》2024年第15期232-235,共4页Journal of Anhui Agricultural Sciences

基　　金：宁夏农林科学院科技创新引导项目“宁夏酿酒葡萄智慧种植关键技术研究与示范”(NKYG-23-02);宁夏自然科学基金项目“基于机器视觉的葡萄免碰采摘路径规划方法研究”(2023AAC03408)。

摘　　要：酿酒葡萄叶幕厚度是园区管理、整形修剪、变量喷药等作业的重要参考依据,但是监测叶幕厚度存在作业量大、效率低、劳动强度高等问题。以贺兰山东麓酿酒葡萄叶幕为目标,基于机器视觉技术研究叶幕厚度的快速估算方法并进行验证。采集葡萄叶幕双目图像及参数,基于HSV颜色空间和形态学处理方法提取左目图像叶幕部分,将提取的图像与右目图像进行校正,采用BM算法对校正后的图像进行立体匹配,计算叶幕范围内所有像素点的三维坐标,依据z值平均值和园区参数计算叶幕厚度估算值。经测试,提出的叶幕厚度估算方法计算结果与实际测量值基本一致,满足酿酒葡萄园实际管理应用监测需求,可为研发叶幕厚度快速测量设备、推进园区管理精细化提供一定的理论基础。The canopy thickness of grapevine is an important reference for vineyards management,shaping and pruning,variable spraying and other operations.However,monitoring the canopy thickness of grapevine has problems,such as high workload,low efficiency and high labor intensity.We focused on the canopy thickness of grapevine from the east Helan Mountain area,and used machine vision technology to study and validate a rapid estimation method for canopy thickness.We collected binocular images and parameters of canopy,extracted the canopy part of the left image based on HSV color space and morphological processing methods,corrected the extracted image with the right eye image,used BM algorithm to perform stereo matching on the corrected image,calculated the three-dimensional coordinates of all pixel points within the leaf canopy range,and calculated the estimated canopy thickness of grapevine based on the average z value and fixed parameters.After testing,the calculation results of the thickness estimation method proposed were basically consistent with the actual measurement values.It could provide a certain theoretical basis for the development of thickness measurement equipment and promoting the refinement of park management.

关 键 词：酿酒葡萄 机器视觉 叶幕厚度 立体匹配 

分 类 号：S126[农业科学—农业基础科学]