Use of deep learning techniques for identification of plant leaf stresses: A review

• Published in: Sustainable computing informatics and systems 2020-12, Vol.28, p.100443, Article 100443
• Main Authors: Noon, Serosh Karim, Amjad, Muhammad, Qureshi, Muhammad Ali, Mannan, Abdul
• Format: Article
• Language: English
• Subjects: Agriculture; Deep networks; Image processing; Plant stress
• ISSN: 2210-5379
• DOI: 10.1016/j.suscom.2020.100443

•A broad survey of 45 recently proposed deep learning techniques is presented by dividing the published works in vegetable, fruits and miscellaneous crops.•Leaf stresses related to 33 different crops using 14 deep learning based models were discussed.•Based on the trends observed in recent literature, some future directions are suggested for researchers and beginners in the field. The use of deep networks in agriculture has increased enormously in the last decade including their use to classify different plant leaf stresses. More recently, a large number of deep learning-based approaches for plant leaf stress identification have been proposed in literature but there are only a few partial efforts to summarize different contributions. Hence, there is a dire need of a detailed survey compiling techniques used for identification of leaf stresses found in a variety of plants. This work presents a review of 45 deep learning-based techniques recently proposed for 33 different crops using 14 famous convolutional neural network architectures. The techniques reviewed were divided in vegetables, fruits and other crops on the basis of stress type, size of dataset, training/test size and the deep network used. The effort will facilitate researchers especially those who are new in this field to get a quick introduction of the trend on using deep learning in plant leaf stress identification.