Journal of Biosystems Engineering
- 제24권1호
- /
- Pages.59-66
- /
- 1999
- /
- 1738-1266(pISSN)
- /
- 2234-1862(eISSN)
기계시각을 이용한 잡초 식별
Weed Identification Using Machine Vision
초록
Weed identification is important for precision farming. A machine vision system was applied to detect weeds. Shape features were analyzed with the binary images obtained from color images of radish, purslane, goosefoot, and crabgrass. Features studied were aspect, roundness, compactness, elongation, PTB, LTP, LTW, and PTAL of each plant. Discriminant analysis was used to classify plant species. The best shape features that distinguished crabgrass were LTP and LTW which distinguished the crabgrass from the others with 100%. Two dimensional discrimination by using LTP and PTB appeared to be effective for distinguishing radish, purslane, and goosefoot.
키워드