2012 Fiscal Year Final Research Report
Web-based support system for diagnosis of rice diseases using image characteristic analysis
Project/Area Number |
22580294
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Single-year Grants |
Section | 一般 |
Research Field |
Agricultural information engineering
|
Research Institution | Hirosaki University |
Principal Investigator |
|
Project Period (FY) |
2010 – 2012
|
Keywords | 病害虫防除 / Web システム / 画像特徴解析 / 病状診断 / EM アルゴリズム / サポートベクターマシン |
Research Abstract |
This study was carried out for the purpose of the method establishment of the image characteristic analysis and the disease judgment for development of the Web-based support system on diagnosis of the rice disease for precision agriculture. Two-dimensional distribution among the color characteristic index HSL at the lesion was analyzed by EM algorithm, and GMM parameters were calculated. As a result that the discrimination experiment between three kinds of main diseases with 9 stages was carried out by SVM discriminant analysis, a right discrimination rate more than an average of 90% was obtained between the different types of diseases. Furthermore, the use experiment of the Web system (prototype) was conducted from handheld device at the farm, and some improvement for practical use was considered.
|