Rice, a staple for billions of people is a major economic crop cultivated all over the world. In Nepal, increased rice production is closely related not just with the national economic health but also with less hunger, better nutrition, lower levels of poverty and a better quality of life. Throughout the life cycle, a rice plant is under constant threat of bacteria, viruses, fungi, nematode, etc which can bring severe diseases like Leaf Blast, Neck Blast, Yellow dwarf etc resulting in reduction of rice production. Agriculture specialist use their naked eyes and experience to identify diseases and the sample may be taken to laboratory for chemical tests. Such traditional way to con- trol rice disease outbreak is time consuming and costly. We propose a highly accurate real-time rice disease identification system which detects rice diseases and provides farmers with necessary advice and strategic decisions to control rice disease. The proposed system uses deep learning models (Deep Convolutional Neural Networks) to detect rice diseases and uses Flutter as a deployment platform to publish web and mobile application. Two different deep learning architecture for object detection (Faster R-CNN and YOLOv3) are compared and a single best architecture is recom- mended for rice disease detection task.
- Ashish Tiwari https://aashishtiwari.com.np
- Nischal Lal Shrestha https://nischal.info.np
- Poshan Pandey https://poshan.com.np
- Rejina Giri https://2017rezina.wordpress.com