[1]汪 晨,张辉辉,乐继旺,等.基于深度学习和遥感影像的松材线虫病疫松树目标检测[J].南京师大学报(自然科学版),2021,44(03):84-89.[doi:10.3969/j.issn.1001-4616.2021.03.013]
　Wang Chen,Zhang Huihui,Le Jiwang,et al.Object Detection to the Pine Trees Affected by Pine Wilt Diseasein Remote Sensing Images Using Deep Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2021,44(03):84-89.[doi:10.3969/j.issn.1001-4616.2021.03.013]
点击复制

基于深度学习和遥感影像的松材线虫病疫松树目标检测()
分享到：

《南京师大学报（自然科学版）》[ISSN:1001-4616/CN:32-1239/N]

卷:
第44卷
期数:
2021年03期
页码:
84-89
栏目:
·地理学·
出版日期:
2021-09-15

文章信息/Info

Title:
Object Detection to the Pine Trees Affected by Pine Wilt Diseasein Remote Sensing Images Using Deep Learning
文章编号:
1001-4616(2021)03-0084-06
作者:
汪 晨; 张辉辉; 乐继旺; 赵 帅
义乌市勘测设计研究院,浙江 义乌 322000
Author(s):
Wang Chen; Zhang Huihui; Le Jiwang; Zhao Shuai
Yiwu Surveying and Design Institute,Yiwu 322000,China
关键词:
松材线虫病; 遥感影像; 深度学习; YOLOv3; 聚类统计; NMS算法
Keywords:
pine wilt disease; remote sensing images; deep learning; YOLOv3; clustering statistics; NMS algorithm
分类号:
P208,P237
DOI:
10.3969/j.issn.1001-4616.2021.03.013
文献标志码:
A
摘要:
针对传统的松材线虫病大范围监测方法耗费高、效率低的问题,研究基于高分辨率无人机遥感影像,采用深度学习方法对病疫区染病松树进行目标检测及地理定位. 建立病疫松树样本数据集,采用YOLO目标检测模型,通过锚框尺寸重算、模型迁移学习等方法进行样本训练. 将大幅影像进行滑窗分割、逐个检测、NMS重叠处理、坐标转换之后,得到研究区内染病松树的数量及其精确坐标. 实验结果表明,本研究算法的准确率为 84.8%,召回率为81.7%. 本研究算法相比传统的目视解译方法,精度接近,但耗时仅为目视解译方法的1/4,更能满足管理部门对松材线虫病害防治大范围、高精度、快速识别的要求.
Abstract:
Aiming at the problems of high cost and low efficiency of the traditional large-scale monitoring method for pine wilt disease,the study is based on high resolution UAV remote sensing images and deep learning methods for target detection and geographic location of diseased pine trees in the epidemic area. Establish a sample data set of diseased pine trees,use the YOLO target detection model,and perform sample training through methods such as anchor frame size recalculation and model migration learning. Divide large images into sliding window segmentation and detect them one by one,after NMS overlap processing,coordinate conversion,the study gets the exact number and coordinates of deseased pine trees in the study area. The result shows that the model to detecting the deseased pine trees achieves a high level of 84.8% of precision rate and 81.7% of recall rate. Compared with the traditional visual interpretation method,the accuracy of this research algorithm is close,but the time-consuming is only 1/4 of the visual interpretation method. It can better meet the requirements of management departments for large-scale,high-precision,and rapid identification of pine wilt disease prevention and control.

参考文献/References:

[1] 张田. 福建省三明市松材线虫病遥感监测预测研究[D]. 北京:北京林业大学,2010.
[2]魏健. 武夷山风景名胜区松材线虫监测技术与防范研究[D]. 福州:福建农林大学,2010.
[3]宁眺,方宇凌,汤坚,等. 松材线虫及其传媒松墨天牛的监测和防治现状[J]. 昆虫知识,2005,42(3):264-269.
[4]邓世晴. 星机地协同的松材线虫病疫区枯死松树监测方法研究[D]. 南昌:东华理工大学,2019.
[5]石进,马盛安,蒋丽雅,等. 航空遥感技术监测松材线虫病的应用[J]. 中国森林病虫,2006,25(1):18-20.
[6]王震,张晓丽,安树杰. 松材线虫病危害的马尾松林木光谱特征分析[J]. 遥感技术与应用,2007,22(3):367-370.
[7]黄宝华. 无人机搭载多光谱相机监测松材线虫病的研究[J]. 广西林业科学,2020,49(3):380-384.
[8]刘遐龄,程多祥,李涛,等. 无人机遥感影像的松材线虫病危害木自动监测技术初探[J]. 中国森林病虫,2019,37(5):16-21.
[9]尹宝才,王文通,王立春. 深度学习研究综述[J]. 北京工业大学学报,2015,41(1):48-59.
[10]CHOLLET F. Deep learning with python[M]. New York:Manning Publications,2017:6-8.
[11]龚健雅. 人工智能时代测绘遥感技术的发展机遇与挑战[J]. 武汉大学学报(信息科学版),2018,43(12):1788-1796.
[12]龚健雅,季顺平. 从摄影测量到计算机视觉[J]. 武汉大学学报(信息科学版),2017,42(11):21-25,118.
[13]张慧,王坤峰,王飞跃. 深度学习在目标视觉检测中的应用进展与展望[J]. 自动化学报,2017,43(8):1289-1305.
[14]朱润生. 基于深度学习的目标检测研究[D]. 北京:北京交通大学,2018.
[15]李旭冬,叶茂,李涛. 基于卷积神经网络的目标检测研究综述[J]. 计算机应用研究,2017,34(10):2881-2886.
[16]REDMON J, FARHADI A. YOLOv3:an incremental improvement[EB/OL]. [2018-04-08]. http://arxiv.org/abs/1804.02767.
[17]AMMAR A,KOUBAA A,AHMED M,et al. Aerial images processing for car detection using convolutional neural networks: comparison between faster R-CNN and YOLOv3[EB/OL]. [2020-04-22]. http://arxiv.org/abs/1910.07234.
[18]李淑敏,冯权泷,梁其椿,等. 基于深度学习的国产高分遥感影像飞机目标自动检测[J]. 遥感技术与应用,2018,33(6):1095-1102.
[19]任嘉锋,熊卫华,吴之昊,等. 基于改进YOLOv3的火灾检测与识别[J]. 计算机系统应用,2019,28(12):175-180.
[20]刘博,王胜正,赵建森,等. 基于Darknet网络和YOLOv3算法的船舶跟踪识别[J]. 计算机应用,2019,39(6):1663-1668.
[21]REDMON J,DIVVALA S,GIRSHICK R,et al. You only look once:unified,real-time object detection[EB/OL]. [2016-05-09]. http://arxiv.org/abs/1506.02640.
[22]REDMON J,FARHADI A. YOLO9000:better,faster,stronger[EB/OL]. [2016-12-25]. http://arxiv.org/abs/1612.08242.
[23]张希盛. 义乌市植物志[M]. 北京:中国林业出版社,2010.

相似文献/References:

[1]丁凯孟,朱长青,罗 文,等.基于自适应PCNN与PCA的遥感影像感知哈希认证算法[J].南京师大学报(自然科学版),2019,42(02):17.[doi:10.3969/j.issn.1001-4616.2019.02.003]
　Ding Kaimeng,Zhu Changqing,Luo Wen,et al.Perceptual Hash Algorithm Based on Adaptive PCNNand PCA for Remote Sensing Image Authentication[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(03):17.[doi:10.3969/j.issn.1001-4616.2019.02.003]

备注/Memo

备注/Memo:
收稿日期:2021-04-08.
基金项目:义乌市自然资源和规划局资助项目(ZXZJZC2019376GK).
通讯作者:汪晨,硕士,工程师,研究方向:WebGIS开发和空间数据可视化、地理国情监测、深度学习. E-mail:ywgi_wangchen@qq.com
更新日期/Last Update: 2021-09-15