Real-time weed detection in rice fields in the Vietnamese Mekong Delta
2020
Nguyen, T.T. (Hokkaido University, Sapporo, Hokkaido (Japan). Graduate School of Agriculture) | Ospina, R. | Noguchi, N. | Okamoto, H. | Ngo, Q.H.
This study introduces an image processing method capable of performing real-time detection of two kinds of weeds in the rice fields of the Vietnamese Mekong Delta (VMD). Two image processing methods were applied and compared in this research: Faster region-based convolutional neural network (R-CNN) and bounding blob analysis. The input images were recorded using a red, green, and blue (RGB) camera. The weeds detection accuracy and processing time were estimated for each method using the same image source data from Vietnam. Both methods were able to detect narrow-leaf and broadleaf weeds on the weed post-emergence stage under uncontrolled light conditions in the rice fields. The results show that bounding blob analysis is simple but effective, with a shorter processing time and higher accuracy than Faster R-CNN.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par Agriculture, Forestry and Fisheries Research Information Technology Center
Découvrez la collection de ce fournisseur de données dans AGRIS