论文摘要
Peach is the fourth largest fruit in China. Peach planting area in China ranks first in the world. In order to improve the utilization rate of pesticides in peach plant protection and enhance the technology of intelligent and accurate target application, human visual characteristics are used as bionic object in this study. Tree crown data is collected through Intel RealSense Depth Camera D435, which will reconstruct the 3 D model of peach tree and provide a foundation for precise and on-demand pesticides application of peach. Firstly, human eye is acted as visual bionic object to study the mechanism of human visual attention and establish a visual bionic model. Then, the depth image and RGB image of peach tree are acquired by Intel RealSense D435, and the interest region is segmented from RGB image. The brightness, direction and color features of the image are extracted by linear filtering, and a Gauss pyramid is built based on multi-resolution analysis method. Local visual contrast is calculated by "center-periphery" operator, and local saliency image is obtained based on normalized combination of normalized operator fusion. The interest region is obtained by segmenting the local saliency image based on the region splitting and merging algorithm. The simulation model of peach tree is reconstructed by matching the segmented image with the depth image based on Intel official SDK. The results showed that the established tree crown model could well retain the 3D morphological characteristics of the tree crown and could provide data support for the target and decision of precise pesticides application.
论文目录
文章来源
类型: 国际会议
作者: Jiangtao Qi,Hongli Tian,Huibin Sun,Yang Li,Hang Liu,Honglei Jia
来源: 第六届国际仿生工程学术大会暨2019年自然启迪科技国际研讨会 2019-09-23
年度: 2019
分类: 基础科学,农业科技,信息科技
专业: 生物学,农业基础科学,园艺,计算机软件及计算机应用
单位: Key Laboratory of Bionic Engineering, Ministry of Education, Jilin University,School of Biological and Agricultural Engineering, Jilin University,Graduate School of agriculture, Kyoto University
分类号: Q811;S662.1;S126
DOI: 10.26914/c.cnkihy.2019.063982
页码: 367
总页数: 1
文件大小: 57k
下载量: 4