日本地球惑星科学連合2022年大会

講演情報

[E] 口頭発表

セッション記号 M (領域外・複数領域) » M-SD 宇宙開発・地球観測

[M-SD41] Geospatial applications for natural resources, environment and agriculture

2022年5月27日(金) 09:00 〜 10:30 202 (幕張メッセ国際会議場)

コンビーナ:Mohamed Shariff Abdul Rashid Bin(Universiti Putra Malaysia )、コンビーナ:高橋 幸弘(北海道大学・大学院理学院・宇宙理学専攻)、座長:高橋 幸弘(北海道大学・大学院理学院・宇宙理学専攻)、Abdul Rashid Bin Mohamed Shariff(Universiti Putra Malaysia)

09:15 〜 09:30

[MSD41-02] Construction of spectral library using handy-spectrometer on the ground for remote-sensing

*高橋 幸弘1成瀬 延康2ゾリゴ ガリッド1砂崎 浩二1石田 百合乃1 (1.北海道大学・大学院理学院・宇宙理学専攻、2.滋賀医科大学・医学部)

キーワード:スペクトル、植生、ライブラリー、リモートセンシング

The detailed reflection spectrum contains a rich information. In particular, for vegetation, spectral measurement is very effective for obtaining highly reliable information on species identification, growth stage, and disease status, which was difficult with conventional broadband observations. However, the spread of hyperspectral observations is limited. There are two main reasons for this. First, many of the traditional hyperspectral sensors, using greating, are large, heavy and expensive. Secondly, the reflection spectrum changes greatly depending on the relationship between the incident angle of sunlight and the angle of view of the sensor, and if these conditions are not met, highly accurate results cannot be obtained. To overcome the latter issue, we decided to construct a spectral library for every combination of angles for various targets. To this end, we devised a small spectroscope that is inexpensive and can automatically record the observation angle and other information, and asked agricultural workers who are not experts in optical measurement to measure the spectrum of crops at high frequency. As a result, we were able to establish a method for efficiently collecting spectral information with sufficient accuracy. In the last year, a total of 200,000 spectral data were recorded in more than 30 fields, and from the analysis, it became clear that the growth stage, disease damage, grass species identifcation, etc. can be performed. By using such a spectral library, we believe that wide-area observations by satellites and drones can achieve high reliability.