Volunteered Geographic Information for Accuracy Assessment of Land Cover Classification Using Remotely Sensed Imagery

View Statistics Email Alert RSS Feed

Information

Details

Project title

Code/計畫編號

MOST105-2625-M865-003

Translated Name/計畫中文名

山區聚落坡地災害風險評估及植生復育之研究-子計畫:自願性地理資訊於衛星影像分類精確度評估之研究(I)

Project Coordinator/計畫主持人

Yuan-Fong Su

Funding Organization/主管機關

National Science and Technology Center for Disaster Reduction

Department/Unit

National Science and Technology Center for Disaster Reduction

Website

https://www.grb.gov.tw/search/planDetail?id=11883792

Year

2016

Start date/計畫起

01-08-2016

Expected Completion/計畫迄

31-07-2017

Bugetid/研究經費

465千元

ResearchField/研究領域

防災工程

本計晝依據坡地災害領域(學門代碼：M1720)研究課題2-1「坡地災害風險管理系統建立暨案例探討」之研究内容1「大量坡地災害(含目錄、潛勢、監測、災損等)收集、補充建置，資訊系統建立、及綜合分析」研提計晝書。為了解廣域的坡地災害潛勢，衛星遙測影像是一項合適的工具。然而衛星影像應用於山區土地利用分類時，分類結果需要足夠且高品質(gold quality) 的驗證資料進行分類精確度的評估，這樣的驗證資料的取得是困難的，故降低了衛星影像土地利用分類結果的可靠度。由於智慧型裝置的普及與社群網路的發展，自願性地理資訊 (Volunteered Geographic Information, VGI)的應用在各領域之中快速地發展，然而VGI的關鍵問題在於本身的資料品質(quality)，尤其是當提供這些資訊的自願者並非受過資料蒐集與整理的專業訓練時，VGI的資料品質則決定了後續應用的廣度與深度，因此極需建立一套評估VGI 可靠度(reliability)的方法，方可進一步應用於衛星影像土地利用分類精確度評估。本計晝擬分三年期，詳盡了解國内外使用VGI於影像分類精確度評估之研究現況、蒐集適用於山區影像土地利用分類之VGI資訊與多元衛星影像，建置一套評估VGI自身可靠度(reliability)與應用於影像分類精確度(accuracy)之評估方法，最後將研究成果於公布於成果網頁。 emote sensing is a state-of-the-art tool for slopeland disaster potential assessment. However, when using remote sensing imagery for land cover mapping in mountainous area, there is a need for validation dataset with gold data quality which is usually difficult to collect. Therefore, it hampers the reliability of land cover mapping results derived from remotely sensed imagery. Due to the rapid development of smart phone and social media, it is ease to acquire and share geographic information through internet and this phenomenon is called Volunteered Geographic Information (VGI). The key issue of VGI is its data quality as these VGI are mostly collected and shared by untrained volunteers. Therefore, the data quality of VGI determines the extent of further applications and there is a need to develop a procedure about the data quality assessment on VGI. With the examined VGI, it is possible to further improve the accuracy of land cover mapping derived from remotely sensed imagery. This project plans a three-year period in which we will review thoroughly the current research status of VGI for land cover mapping, and will collect various types of remotely sensed imagery for land cover mapping in mountainous area, and develop a procedure for reliability of VGI and accuracy of land cover mapping derived from remotely sensed imagery. Finally the results of this project will publish on webpage and feedback to international VGI project such as Geo-Wiki.

DSpace CRIS

Volunteered Geographic Information for Accuracy Assessment of Land Cover Classification Using Remotely Sensed Imagery

Details

Description