Development of remote sensing for plant stress detection

Abstract

Early detection of plant stress is important in agriculture world in order to minimize crop losses. Previously, plant stress detection was done manually, so there’s a limitation in manpower ability. There are various methods to detect plant stress. This project will detect plant stress by using image processing method. Application of image segmentation, threshold, feature extraction, colours comparison and colours recognition using HSV in Matlab software determined which Harumanis leaves that have been affected. There will be only showing several detection of plant stress since this project is only focus on certain stress factors. However, problems and constraints have been pointed out such as shape of leaf taken in different angle and distance, the resolution of image captured and various colour of leaf. Yellow would give significant to the lack of nutrients, while brown is a result of over fertilizing. As for white and black, it shows that Harumanis leaf have been attacked by pest or some other virus. As a result for this project, 80 % of the selected images are recognized as plant stress. The experiments were performed by proposed methods and results were collected and analyzed.