Real-Time Detection and Instance Segmentation of Strawberry in Unstructured Environment

作     者：Chengjun Wang Fan Ding Yiwen Wang Renyuan Wu Xingyu Yao Chengjie Jiang Liuyi Ling 

作者机构：School of Artificial IntelligenceAnhui University of Science and TechnologyHuainan232001China School of Mechanical EngineeringAnhui University of Science and TechnologyHuainan232001China 

出 版 物：《Computers, Materials & Continua》 (计算机、材料和连续体（英文）)

年 卷 期：2024年第78卷第1期

页      面：1481-1501页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：funded by Anhui Provincial Natural Science Foundation(No.2208085ME128) the Anhui University-Level Special Project of Anhui University of Science and Technology(No.XCZX2021-01) the Research and the Development Fund of the Institute of Environmental Friendly Materials and Occupational Health,Anhui University of Science and Technology(No.ALW2022YF06) Anhui Province New Era Education Quality Project(Graduate Education)(No.2022xscx073) 

主　　题：YOLACT real-time detection instance segmentation attention mechanism strawberry 

摘      要：The real-time detection and instance segmentation of strawberries constitute fundamental components in the development of strawberry harvesting ***-time identification of strawberries in an unstructured envi-ronment is a challenging *** instance segmentation algorithms for strawberries suffer from issues such as poor real-time performance and low *** this end,the present study proposes an Efficient YOLACT(E-YOLACT)algorithm for strawberry detection and segmentation based on the YOLACT *** key enhancements of the E-YOLACT encompass the development of a lightweight attention mechanism,pyramid squeeze shuffle attention(PSSA),for efficient feature ***,an attention-guided context-feature pyramid network(AC-FPN)is employed instead of FPN to optimize the architecture’s ***,a feature-enhanced model(FEM)is introduced to enhance the prediction head’s capabilities,while efficient fast non-maximum suppression(EF-NMS)is devised to improve non-maximum *** experimental results demonstrate that the E-YOLACT achieves a Box-mAP and Mask-mAP of 77.9 and 76.6,respectively,on the custom ***,it exhibits an impressive category accuracy of 93.5%.Notably,the E-YOLACT also demonstrates a remarkable real-time detection capability with a speed of 34.8 *** method proposed in this article presents an efficient approach for the vision system of a strawberry-picking robot.