PLANT DISEASE DETECTION USING MACHINE LEARNING

Abstract

With the ever-growing population of human beings and animals, the demand for the crops to sustain them has also been increasing. We have tried many methods to increase the yields of the crops such as the introduction of HYV (High Yielding Varieties), increasing the area under cultivation and many more. One such problem to the cultivation is the plant-based diseases caused by various pathogens. In general, cultivated plants are much more susceptible to plant diseases because of the proximity of similar plants (sometimes hundreds of kms) which causes the pathogens to spread more freely. Generally, plant diseases are classified into two categories- Infectious and Noninfectious. Infectious plant diseases are caused by organisms like viruses, bacteria, mycoplasma, fungus, etc. Non Infectious plant diseases are caused mainly due to environmental factors such as unfavorable moisture, sunlight, temperature, toxic elements in the environment, insufficient minerals in the soil, etc. Proper diagnosis of such diseases is very important to stop the spread of the diseases and save the cultivation. A common way to diagnose such diseases is based on analyzing the appearance of the leaves, stem, and roots for any abnormalities like discoloration, overdevelopment or underdevelopment, misshape, holes, swollen, etc. As estimated by the UN about 80% of farmers in developing countries are smallholder farmers so the protection of plant diseases becomes very crucial in saving their livelihoods and achieving food security. Various organizations are engaged in providing information to the farmers online and through various programs. As the internet is widely available and the computation power of modern technology is capable enough, the diagnosis of plant disease with the help of machine learning and computer vision will be very helpful to the farmers. As more than half of the world population has access to smartphones containing cameras and internet capabilities along with good processing powers, so the development of image-based plant disease detection will be very helpful and beneficial for the smallholder farmers and society. In this project, we will look at the feasibility of developing such a platform by looking at the detection of plant diseases through machine learning techniques.