{"title":"APNet-YOLOv8s: A real-time automatic aquatic plants recognition algorithm for complex environments","authors":"","doi":"10.1016/j.ecolind.2024.112597","DOIUrl":null,"url":null,"abstract":"<div><p>Deep learning techniques have been widely utilized for image recognition tasks. However, these techniques remain challenging in detecting aquatic plants due to their complex growing environments, long phenological periods, high species similarity, and the fact that they are often obscured by surrounding objects. To overcome these challenges, this study presents a comprehensive dataset of aquatic plant images in complex environments (DS-AP) and proposes a novel method, APNet-YOLOv8s. APNet-YOLOv8s integrates three modules: the Global Receptive Field-Space Pooling Pyramid-Fast (GRF-SPPF), the Shuffle Attention (SA) Mechanism, and the Fast Detection (FD), each designed to tackle specific challenges in aquatic plant detection. The performance of APNet-YOLOv8s was thoroughly evaluated using the DS-AP dataset. The results demonstrate that APNet-YOLOv8s significantly outperforms YOLOv8s, achieving a mean average precision (mAP50) of 75.3 % with a 2.7 % improvement, and a frame per second (FPS) rate of 30.5 with a 50.2 % increase. Moreover, APNet-YOLOv8s accurately and rapidly identifies aquatic plants in Gradient-weighted Class Activation Mapping (Grad-CAM) visualizations and real-world scenarios, highlighting its practical applications in complex environments. Overall, this study advances the application of deep learning in aquatic environments, providing a potential solution for rapid detection in other challenging environments.</p></div>","PeriodicalId":11459,"journal":{"name":"Ecological Indicators","volume":null,"pages":null},"PeriodicalIF":7.0000,"publicationDate":"2024-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1470160X24010549/pdfft?md5=66d29df73c989a4d8a27e1364ebadeb2&pid=1-s2.0-S1470160X24010549-main.pdf","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Ecological Indicators","FirstCategoryId":"93","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1470160X24010549","RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0
Abstract
Deep learning techniques have been widely utilized for image recognition tasks. However, these techniques remain challenging in detecting aquatic plants due to their complex growing environments, long phenological periods, high species similarity, and the fact that they are often obscured by surrounding objects. To overcome these challenges, this study presents a comprehensive dataset of aquatic plant images in complex environments (DS-AP) and proposes a novel method, APNet-YOLOv8s. APNet-YOLOv8s integrates three modules: the Global Receptive Field-Space Pooling Pyramid-Fast (GRF-SPPF), the Shuffle Attention (SA) Mechanism, and the Fast Detection (FD), each designed to tackle specific challenges in aquatic plant detection. The performance of APNet-YOLOv8s was thoroughly evaluated using the DS-AP dataset. The results demonstrate that APNet-YOLOv8s significantly outperforms YOLOv8s, achieving a mean average precision (mAP50) of 75.3 % with a 2.7 % improvement, and a frame per second (FPS) rate of 30.5 with a 50.2 % increase. Moreover, APNet-YOLOv8s accurately and rapidly identifies aquatic plants in Gradient-weighted Class Activation Mapping (Grad-CAM) visualizations and real-world scenarios, highlighting its practical applications in complex environments. Overall, this study advances the application of deep learning in aquatic environments, providing a potential solution for rapid detection in other challenging environments.
期刊介绍:
The ultimate aim of Ecological Indicators is to integrate the monitoring and assessment of ecological and environmental indicators with management practices. The journal provides a forum for the discussion of the applied scientific development and review of traditional indicator approaches as well as for theoretical, modelling and quantitative applications such as index development. Research into the following areas will be published.
• All aspects of ecological and environmental indicators and indices.
• New indicators, and new approaches and methods for indicator development, testing and use.
• Development and modelling of indices, e.g. application of indicator suites across multiple scales and resources.
• Analysis and research of resource, system- and scale-specific indicators.
• Methods for integration of social and other valuation metrics for the production of scientifically rigorous and politically-relevant assessments using indicator-based monitoring and assessment programs.
• How research indicators can be transformed into direct application for management purposes.
• Broader assessment objectives and methods, e.g. biodiversity, biological integrity, and sustainability, through the use of indicators.
• Resource-specific indicators such as landscape, agroecosystems, forests, wetlands, etc.