基于机器视觉的天然橡胶树割胶轨迹识别规划研究  

Research on recognition and planning of tapping trajectory of natural rubber tree based on machine vision

作　　者：孙泽瑾 邢洁洁 胡宏男 张喜瑞[1] 董学虎 邓怡国[4] Sun Zejin;Xing Jiejie;Hu Hongnan;Zhang Xirui;Dong Xuehu;Deng Yiguo(Mechanical and Electrical Engineering College,Hainan University,Haikou,570228,China;College of Mechanical and Electrical Engineering,Zhongkai University of Agriculture and Engineering,Guangzhou,510225,China;Hainan Agricultural Machinery Appraisal and Promotion Station,Haikou,570312,China;Tropical Agricultural Machinery Research Institute,Chinese Academy of Tropical Agricultural Sciences,Zhanjiang,524091,China)

机构地区：[1]海南大学机电工程学院,海口市570228 [2]仲恺农业工程学院机电工程学院,广州市510225 [3]海南省农业机械鉴定推广站,海口市570312 [4]中国热带农业科学院农业机械研究所,广东湛江524091

出　　处：《中国农机化学报》2022年第5期102-108,共7页Journal of Chinese Agricultural Mechanization

基　　金：海南省院士创新平台科研专项(YSPTZX202008);国家天然橡胶产业技术体系(CARS—33—JX2);海南省自然科学基金青年基金项目(520QN233);海南省院士工作站专项(HD—YSZX—202101)。

摘　　要：针对光照条件复杂多变、目标前景与背景特征相近等外界因素影响割胶轨迹识别规划的难题,提出一种天然橡胶树割胶轨迹识别规划组合优化算法。首先对原始图像进行图像预处理,降低图像噪声影响,增强割胶轮廓与背景的特征信息;然后采用OTSU算法进行阈值分割,获得二值图像,利用形态学开运算消除残余噪声点;再通过对比分析不同边缘提取算子,选用Canny算子提取割胶轨迹;最后采用轨迹拟合方法规划出下一条割胶轨迹。为验证所提算法的稳定性和有效性,对采集于海南国家天然橡胶林的180张图像进行试验。结果表明:所提出的组合优化算法在自然光照条件下可识别规划出下一条割胶轨迹,且识别规划成功率达89.4%,可为研发智能割胶机器人提供技术参考。In view of the external factors that affect the recognition and planning of tapping trajectory,such as complex and variable lighting conditions and similar target and background,a combinatorial optimization algorithm for the recognition and planning of tapping trajectory of natural rubber tree was proposed.Firstly,the algorithm preprocessed the original image to reduce the impact of image noise and to enhance the feature information between the profile and background.Then,OTSU algorithm was used for threshold segmentation to obtain binary image and morphology operation was used to eliminate residual noise points,after which the Canny operator was selected to extract the glue track by comparing and analyzing different boundary extraction operators.Finally,the trajectory fitting operation was used to plan the new tapping trajectory.In order to verify the stability and effectiveness of the proposed algorithm,the experiment was carried out on 180 images collected in Hainan National Natural Rubber Forest.The results showed that the proposed combinatorial optimization algorithm could recognize and plan the next tapping trajectory under natural light conditions and the success rate of recognition and planning reached 89.4%.It can provide technical reference for the research and development of intelligent tapping robot.

关 键 词：天然橡胶 割胶轨迹 识别规划 特征信息 智能割胶 

分 类 号：S126[农业科学—农业基础科学] TP391.4[自动化与计算机技术—计算机应用技术]