Nhận dạng bệnh trên lá lúa bằng phương pháp học chuyển giao

0 Người đánh giá. Xếp hạng trung bình 0

Tác giả: Thái Nghe Nguyễn, Thị Phương Thanh Trương

Ngôn ngữ: Vie

Ký hiệu phân loại:

Thông tin xuất bản: Khoa học (Đại học Cần Thơ) 2022

Mô tả vật lý: 45298

Bộ sưu tập: Metadata

ID: 400200

Thêm vào giỏ Liên kết toàn văn

Nông nghiệp thông minh là xu hướng nghiên cứu và ứng dụng rất được quan tâm gần đây. Phát triển ngành nông nghiệp, đặc biệt là cây lúa, là một trong những lĩnh vực quan trọng trong phát triển kinh tế, xã hội. Cây lúa là một sản phẩm có lợi thế lớn của vùng đồng bằng sông Cửu Long, nhưng khó khăn trong trồng lúa là việc xuất hiện những loại bệnh như đạo ôn, đốm nâu, cháy bìa lá và bọ gai đã làm giảm cả sản lượng và chất lượng của việc trồng lúa. Vì vậy, việc phát hiện các bệnh phổ biến trên cây lúa nhằm giúp người dân nâng cao năng suất là vấn đề cấp thiết. Nghiên cứu này đề xuất giải pháp trong nhận dạng bệnh hại trên lá lúa bằng mô hình học sâu. Mô hình đã ứng dụng việc học chuyển giao với mô hình học sâu Inception V3 để phân lớp một số loại bệnh phổ biến trên lá lúa. Thực nghiệm trên tập dữ liệu thu thập từ các nguồn khác nhau bao gồm 2.500 hình ảnh cho thấy mô hình đạt độ chính xác 97,4%. Kết quả này rất khả thi để ứng dụng vào thực tế nhằm dự đoán các loại bệnh trên lá lúa thông qua ảnh chụp, từ đó đề xuất giải pháp phòng trị phù hợp giúp người dân nâng cao năng suất trồng lúa., Tóm tắt tiếng anh, Smart agriculture is a research and application trend which is receiving high attention. The development of agriculture, especially rice, is one of the important sector in socio-economic development. Rice is a product with great advantages of the Mekong Delta region, however, the difficulty in rice cultivation is the occurrence of diseases such as rice blast, brown spot, leaf blight and thorny beetle that causes reduction in both yield and quality of rice cultivation. Therefore, the detection of common diseases on rice in order to help farmers to improve productivity is an urgent need. This study proposes a solution to identify diseases on rice leaves by deep learning model. The model has applied transfer learning with Inception V3 deep learning model to classify some common diseases on rice leaves. Experiments on a dataset of 2,500 images showed that the model had an accuracy of 97.4%. This result is very feasible and applicable to predict diseases on rice leaves through photographs, thereby proposing appropriate prevention and treatment solutions to help farmers improving rice productivity.

1. Đọc trực tuyến

1. Đọc trực tuyến

Tạo bộ sưu tập với mã QR