Multi-Feature Optimization based Plant Leaves Alignment

• Main Authors: Jheng-Yao Lin, 林政堯
• Other Authors: Duan-Yu Chen
• Format: Others
• Language: en_US
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/14068746468767756203

碩士 === 元智大學 === 電機工程學系 === 104 === The technology used to the automation-agriculture and greenhouse is very flourishing. Therefore, we think plant leaves alignment is an important key point to improve the efficacy of those technologies. However, plant leaves alignment is a hard task even if the plant is in the clear environment. If people want to get the high efficacy of leaves alignment, they usually need to use some high accuracy instrument like 3D camera or 3D scanner. Therefore, we proposed the method that used only common camera to reach alignment for multiple potentially overlapping leaves. First, we use the input image taken by common camera and CNN model to get the edge information between leaves of plant. Then, we apply edge information to operate the features to our database. Finally, we apply objective function operation to get the optimal result. The results show that our proposed method can solve the task with overlapping leaves.