首页|基于深度学习的近地面草原鼠洞识别计数关键问题研究与应用

基于深度学习的近地面草原鼠洞识别计数关键问题研究与应用

扫码查看
原文链接
  • 万方数据
  • 维普
鼠洞密度可用于评估草原鼠害发生程度.在近地面鼠洞图片采集与识别中,图像分辨率和拍摄倾角是影响时效性和识别性能的关键因素.为此,设计了带有倾角传感器的图像采集装置,在锡林郭勒草原采集2 325张鼠洞图片并进行手工标注,对比分析了 3种图片输入尺寸(416 pixel ×416 pixel、608 pixel ×608 pixel、1 024 pixel × 1 024 pixel)、4 类拍摄倾角(21°、32°、41°、51°)、2 种目标识别模型(YOLOv3 和 YOLOv4)对识别性能的影响.结果表明:YOLOv4模型在输入图片尺寸为416 pixel ×416 pixel时能取得最优的性能.当拍摄倾角为41°时,识别性能最优;当拍摄倾角为32°时,识别性能最差.与近3年发表的鼠洞识别方法进行对比分析,验证了本文方法的性能先进性.研究结果可为草原鼠害的智能化监测提供技术支撑.
Study on key problems for rat hole recognition and count near ground based on deep learn-ing and its application
;Rat hole density serves as a significant indicator for assessing the extent of rat damage in grasslands.In the image acquisition for rat hole recognition and count near ground by deep learning,the optimal image input size and shooting angle are crucial.To address these considerations,an image acquisition device equipped with an incli-nation sensor was designed,and a dataset of 2 325 rat hole images was collected from Xilingol grassland and manual-ly annotated.The recognition performance was compared under 3 image input sizes(416 pixel ×416 pixel,608 pixel × 608 pixel,1 024 pixel × 1 024 pixel),4 shooting angles(21°,32°,41°,and 51°),and 2 target recognition mod-els(YOLOv3 and YOLOv4).The findings revealed that YOLOv4 achieved superior performance when the image in-put size was set as 416 pixel ×416 pixel.Furthermore,the recognition performance was optimal at the shooting angle of 41°,while it was the poorest at 32°.The proposed method was validated by comparing it with relevant approaches published within the past three years.These results offered valuable insights to support the development of intelligent monitoring technologies for assessing rat damage in grassland ecosystems.

object detectiongrassland ecosystemmachine visionrat holeYOLOv3 modelYOLOv4 model

郭秀明、王大伟、刘升平、诸叶平、刘晓辉、林克剑、王佳宇、李非

展开 >

中国农业科学院农业信息研究所,农业农村部农业信息服务技术重点实验室,北京 100081

中国农业科学院植物保护研究所,植物病虫害综合治理全国重点实验室,北京 100081

中国农业科学院西部农业研究中心,新疆昌吉 831100

中国农业科学院草原研究所,农业农村部人工草地生物灾害监测与绿色防控重点实验室,内蒙古呼和浩特 010010

锡林郭勒盟草原工作站,内蒙古锡林浩特 026000

展开 >

目标检测 草原生态 机器视觉 鼠洞 YOLOv3模型 YOLOv4模型

内蒙古自治区科技计划内蒙古自治区科技计划中国农业科学院平台提质增效项目中国农业科学院创新工程中国农业科学院北方农牧业科技创新中心项目"天池英才"引进计划

2022YFSJ00102020GG0112Y2021PT03CAAS-ASTIP-2016-AIIBFGJ2022007

2024

10.3969/j.issn.1004-1524.20230858

浙江农业学报
浙江省农业科学院 浙江省农学会

浙江农业学报

CSTPCD北大核心
影响因子：0.765
ISSN：1004-1524
年,卷(期)：2024.36(9)