Application of image processing technics and deep learning in field of crops stress

One of the most important challenges of agricultural production growers and food security is environmental stress (biotic and Abiotic) specially caused by global changes. Sustainable yield could be accessible by identifying environmental stresses using physiological studies. Physiological and phenotypical researches on crop have been based on labor-intensive conventional, distractive and time-consumer methods, as laborious and farm tasks, for many years. To address this issue, rapid approaches such as using machine vision technologies, machine learning and deep learning's algorithms, are in urgent. These methods have had positive effect on prediction or identification of stresses by monitoring crop phenotypical and physiological changes. In this paper, the latest image technology, vegetation indices and a diversity of deep learning algorithms involved in plant stress, are reviewed. Furthermore, the most functional algorithms of convolutional neural networks are summarized. On the other hand, the current challenges of application of image processing approaches and artificial intelligent in plant stressed are discussed.