摘要
本文基于HSV颜色模型实现超高分辨率影像单株变色松树的快速识别,并充分利用多光谱卫星影像、激光雷达点云等多源遥感数据与松树冠层图形几何特征,减少枯死草地、阴影、裸地和道路的干扰,并通过目视解译及野外验证对比不同分辨率不同色彩空间影像提取结果。研究证明,激光雷达点云及松树冠层图形几何特征的应用可大幅降低枯死草地、阴影、裸地和道路的干扰,将错分率从60%降低至20%,并且基于HSV颜色模型的识别精度远高于RGB图像;同时,基于空间分辨率为0.04 m与0.1 m的识别结果基本相同。本文为基于无人机高分辨率影像变色松树识别的数据获取提供了方向,为基于多源遥感数据快速高效精准识别变色松树开辟了一条新的技术思路和手段。
This paper is based on the HSV color model to realize the rapid identification of individual red-attack pine trees in ultra-high-resolution images,and makes full use of multi-source remote sensing data such as multi-spectral satellite images,LIDAR point clouds,and geometric features of pine tree canopy graphics to reduce dead grass,shadows,and shadows.The interference of bare land and roads,and through visual interpretation and field verification,compare the extraction results of images with different resolutions and different color spaces.Studies have proved that the application of LIDAR point cloud and pine tree canopy graphic geometric features can greatly reduce the interference of dead grass,shadows,bare land and roads,reducing the error rate from 60%to 20%,and recognition based on the HSV color model.The accuracy is much higher than that of RGB images;at the same time,the recognition results based on the spatial resolution of 0.04 m and 0.1 m are basically the same.This article provides a direction for data acquisition based on the identification of red-attack pine trees based on high-resolution UAV images,and opens up a new technical idea and method for quickly,efficiently and accurately identifying red-attack pine trees based on multi-source remote sensing data.
作者
李琴
夏进亮
余榕
周火生
陈勇明
LI Qin;XIA Jinliang;YU Rong;ZHOU Huosheng;CHEN Yongming(Surveying and Mapping Institute,Lands and Resource Department of Guangdong Province,Guangzhou 510500,China)
出处
《热带地貌》
2021年第1期54-58,共5页
Tropical Geomorphology
基金
广东省自然资源厅科技项目(GDZRZYKJ-ZC2020003)
关键词
HSV颜色模型
多源遥感
变色松树
遥感识别
HSV color model
multi-source remote sensing
red-attack pine trees
remote sensing recognition
作者简介
李琴(1991-),女,贵州织金人,硕士,工程师,主要从事摄影测量遥感与地理信息数据分析研究。E-mail:13618570901@163.com;通信作者:夏进亮(1990-),男,湖北十堰人,硕士,主要从事测绘工程、摄影测量与遥感研究。E-mail:554217680@qq.com
