Development of predicting system of dust emission for farm land

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K07645
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Rural environmental engineering/Planning
• Research Institution: Chiba University
• Principal Investigator: MATSUOKA Nobuhiro 千葉大学, 大学院園芸学研究科, 教授 (80212215)
• Co-Investigator(Kenkyū-buntansha): 間野 正美 千葉大学, 大学院園芸学研究科, 助教 (10391210) ; 木村 玲二 鳥取大学, 乾燥地研究センター, 准教授 (80315457)
• Co-Investigator(Renkei-kenkyūsha): KIMURA Reiji 鳥取大学, 乾燥地研究センター, 准教授 (80315457)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 農業気象・環境学 / 気象災害 / 画像情報処理・画像認識 / 災害リスク評価 / 機械学習 / 風食 / ダスト

Outline of Final Research Achievements

Dust generated from bare ground in dry and wet areas exerts various adverse effects on the lives of the surrounding residents, and development of prediction method and deterrence technology is awaited. We developed a dust emission risk measurement system in Yachimata city, Chiba prefecture. This system consists of dust sensor, anemometer, soil moisture meter, single board computer, multispectral camera.
From the research results, it was possible to correct luminance and color as preprocessing of machine learning, but shadows are formed by irregularities on the soil surface. If they are removed, you can obtain images that are easier to use for machine learning than it was suggested. In order to solve this problem, we found that introduction of SfM algorithm and detection of shadows by thermal infrared image are necessary.

Free Research Field

農業気象・環境学