DISCOVERY OF AGRICULTURAL DISEASES BY DEEP LEARNING AND OBJECT DETECTION


Karakaya M., Çelebi M. F., Gok A. E., Ersoy S.

ENVIRONMENTAL ENGINEERING AND MANAGEMENT JOURNAL, cilt.21, sa.1, ss.163-173, 2022 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 21 Sayı: 1
  • Basım Tarihi: 2022
  • Dergi Adı: ENVIRONMENTAL ENGINEERING AND MANAGEMENT JOURNAL
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aqualine, Aquatic Science & Fisheries Abstracts (ASFA), CAB Abstracts, Environment Index, Greenfile, Pollution Abstracts, Veterinary Science Database
  • Sayfa Sayıları: ss.163-173
  • Anahtar Kelimeler: agricultural disease, deep learning, disease detection, object detection
  • Marmara Üniversitesi Adresli: Evet

Özet

In this study deep learning and object detection models for image-based plant disease recognition have been carried. Trained models were tested on pictures and in real-time with a video camera for five different diseases in tomato leaves. Object detection algorithm was implemented from the personal computer, and deep learning models were applied via Google Colab. Real-time object detection was achieved in the developed model with YOLOv5 algorithm with the highest accuracy of 93.38% in validation accuracy and 94.48% in training accuracy with the highest value of 92.96% in precision. Furthermore, it has been observed that YOLOv5 algorithm gives faster and more accurate results than the previous versions of YOLO.