機載多譜影像系統偵測作物淹水災害之研究

Abstract

本研究目的在測試直昇機機載多光譜影像系絞獲取農業災害資訊的可行性，測試項目包括分兩大類，包括基本的農作物因水災淹沒受損的光譜反應、及機載影像系統的幣合運作測試。作物淹水傷害試驗方面，本研究對小白菜、莧菜、及西瓜實施3至12小時定時淹水，及對小白菜實施長期（48小時）淹水後，利用高光譜輻射儀及多光譜成像儀施測，縱現作物受淹水災害在一定時間內（如1~2天），若暴雨停止及排水得宜，作物仍可依賴其自行調適機制回復正常生長狀態，但長至48小時的淹水，將對作物產生不可回逆轉的傷害，凋萎關始後，再行排水亦無法恢復正常，反應在光譜上則為NDVI值的急速下降。在攝像系統方面，本計畫採用輕型直昇機以為機載系統載具，直昇機雖不如定翼機可獲取高精度航照或以掃描儀攝像，但藉由低設備維護與迅速動員就位的優點，可作為輔助資料的要求。直昇機載系統的研究工作包括儀器測試、設備掛載、系統整合、觀測作業、影像處理分析等。 The study used a hehcopter-mounted muhispectral image system to establish an auxiliary tool for assessing crop inundant damage. A series of flooding test were applied to three types of plants to find the spectrum information of plants under both health and injured conditions. These data are the base to build database for ranking disaster level. The study also established a helicopter-mounted system to assess multispectral crop image above ground. Comparing with other data source, the system provides fast data acquisition ability. Feasibility test of the system including instrument integration, data acquisition procedure, image processing, and image analysis were carried out in the study.