Detection of Onion Leaf Disease Using Hybridized Feature Extraction and Feature Selection Approach

Onion (Allium Cepa) is one of the most important vegetable and commercial plants that is being grown all around the world for more than 3000 years. Just like several other crop plants, Onion plants too can be attacked by pests and diseases of various kind, this attacks do give rise to low yields, bad quality and of course shortages of this important plants. Visual observation and analysis for detection of onion leaf diseases, if handed over to computing, using Machine Learning techniques, is more efficient, fast, cost saving, consistent, more reliable and highly accurate compare to what any human disease-expert eyes can offer. This work makes use of the prepared datasets of onion leaf digital images, after image preprocessing, some features were extracted/selected using Grey Level Co-occurrence Matrix (GLCM) and Particle Swarm Optimization (PSO) algorithms, the selected/extracted features then fed into classifier algorithms for eventual classification into healthy or unhealthy onion leaf.