文献类型：期刊文献

英文题名：Precious Tree Pest Identification with Improved Instance Segmentation Model in Real Complex Natural Environments

作者：Guo, Ying[1] Gao, Junjia[1] Wang, Xuefeng[1] Jia, Hongyan[2] Wang, Yanan[2] Zeng, Yi[2] Tian, Xin[1] Mu, Xiyun[3] Chen, Yan[1] OuYang, Xuan[1]

第一作者：郭颖

通信作者：Wang, XF[1]

机构：[1]Chinese Acad Forestry, Inst Forest Resource Informat Tech, Beijing 100091, Peoples R China;[2]Chinese Acad Forestry, Expt Ctr Trop Forestry CAF, Pingxiang 532600, Peoples R China;[3]Forestry Res Inst Chifeng, Chifeng 024000, Peoples R China

年份：2022

卷号：13

期号：12

外文期刊名：FORESTS

收录：;EI(收录号：20225213314676);Scopus(收录号：2-s2.0-85144599535);WOS:【SCI-EXPANDED(收录号:WOS:000901157300001)】；

基金：This research was funded by the Special Funds for Fundamental Research Business Expenses of the Central Public Welfare Research Institution’s “Study on Image Diagnosis Technology of Main Diseases and Insect Pests of Rare Tree Species”, grant number CAFYBB2021ZB002; “Study on classification of tree species based on cooperative multi optical sensors”, grant number CAFYBB2022SY030; “Study on key technologies of forest resources output”, grant number CAFYBB2021SY006 and the Zhejiang Provincial Academy Cooperative Forestry Science and Technology Project “Research and application of forest resources monitoring technology in zhejiang province based on sky earth integration”, grant number 2020SY02.

语种：英文

外文关键词：precious trees control; small pest segmentation; instance segmentation; Mask RCNN; Swin Transformer

摘要：It is crucial to accurately identify precious tree pests in a real, complex natural environment in order to monitor the growth of precious trees and provide growers with the information they need to make effective decisions. However, pest identification in real complex natural environments is confronted with several obstacles, including a lack of contrast between the pests and the background, the overlapping and occlusion of leaves, numerous variations in pest size and complexity, and a great deal of image noise. The purpose of the study was to construct a segmentation method for identifying precious tree pests in a complex natural environment. The backbone of an existing Mask region-based convolutional neural network was replaced with a Swin Transformer to improve its feature extraction capability. The experimental findings demonstrated that the suggested method successfully segmented pests in a variety of situations, including shaded, overlapped, and foliage- and branch-obscured pests. The proposed method outperformed the two competing methods, indicating that it is capable of accurately segmenting pests in a complex natural environment and provides a solution for achieving accurate segmentation of precious tree pests and long-term automatic growth monitoring.