当前位置: 首页 > 文章 > 基于深度学习的果园道路导航线生成算法研究 湖南农业大学学报(自然科学版) 2019,45 (6) 674-678
Position: Home > Articles > Research on generating algorithm of orchard road navigation line based on deep learning Journal of Hunan Agricultural University(Natural Sciences) 2019,45 (6) 674-678

基于深度学习的果园道路导航线生成算法研究

作  者:
王毅;刘波;熊龙烨;王卓;杨长辉
单  位:
重庆理工大学机械工程学院;重庆大学机械工程学院;西安交通大学机械工程学院
关键词:
果园环境;图像处理;深度学习;导航线
摘  要:
针对目前果园环境视觉导航线提取受光照、杂草干扰,而已有的导航线生成算法过于复杂且适用范围较窄的问题,提出了一种基于深度学习的方法来提取果园道路导航线:采用YOLOV3卷积神经网络提取果园道路图像上的特征点,并通过最小二乘法拟合生成导航线。通过采集的800张果园道路图片作为训练集,进而在240张图片组成的独立测试集上进行测试,总体识别率为95.37%,在杂草较少、杂草较多、高光照和正常光照的环境下,导航线平均偏差分别为2.15、2.28、2.32、2.41像素;平均偏移距离分别为3.4、3.5、2.7、3.6 cm。
译  名:
Research on generating algorithm of orchard road navigation line based on deep learning
作  者:
WANG Yi;LIU Bo;XIONG Longye;WANG Zhuo;YANG Changhui;College of Mechanical Engineering, Chongqing University of Technology;College of Mechanical Engineering, Chongqing University;School of Mechanical Engineering, Xi'an Jiaotong University;
关键词:
orchard environment;;image processing;;deep learning;;navigation line
摘  要:
In view of the fact that the current orchard environment visual navigation line is susceptible to interference from light and weeds, and the existing navigation line generation algorithm is too complicated and has a narrow application range, this paper proposes a method to extract orchard road navigation line based on deep learning. YOLOV3 convolutional neural network is used to extract the feature points on the image and generate the navigation line by least square fitting. 800 pictures of orchard roads collected under different conditions are used as training sets, and then tested on an independent test set consisting of 240 pictures. The overall recognition rate is 95.37%. In the environment with fewer weeds, more weeds, high light and normal light, the mean deviation of navigation line is 2.15 pixels, 2.28 pixels, 2.32 pixels and 2.41 pixels, and the mean deviation distance is 3.4 cm, 3.5 cm, 2.7 cm and 3.6 cm, respectively.

相似文章

计量
文章访问数: 10
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊