Executive Summary : | Sugarcane is the primary sugar crop of India that play an important role in domestic as well as export market. Sugarcane cultivation promotes employment and livelihood to more than 7.5% of rural population. Nematodes are one of the major constrain for successful sugarcane farming. The root damage caused by nematodes leads to 10-40% loss of sugarcane yield. The root-knot nematode (Meloidogyne incognita), lesion nematode (Pratylenchus spp), lance nematode (Hoplolaimus spp), stunt nematode (Tylenchorhynchus spp) and spiral nematode (Helicotylenchus spp) are the major nematode pest occurring in sugarcane fields of India. Timely detection, diagnosis and monitoring of these nematode pests is very essential to avoid damages caused by them. During monitoring and diagnosis, the soil samples collected from the fields are processed by sieving and decanting method that result suspension containing nematodes. The suspension needs to be placed in a nematode counting dish and viewed through microscopes. The identification and counting job is done by trained nematode taxonomist who identify them based on the morphological structures like head, stylet, body length and etc. Since the suspension obtained from soil sample ought to have different nematodes like plant parasite, bacterial feeder, fungal feeder, predator and insect parasite, diagnosis can be done by the experts only. Detection and counting of target nematode is a challenging task due to the reasons like require trained nematode taxonomist, intra-rater and inter-rater variability issues and a time consuming process. The need of the day is an automation of nematode detection and diagnosis process. Once the nematodes diagnosed in time, the farmers can avoid loss due to nematodes. Hence, this proposal is formulated with an objective to develop a computer vision diagnostic kit through novel deep learning- artificial intelligence app or algorithm for accurate identification and quantification of key nematode pests of sugarcane. As a first step, deep learning AI algorithms will be developed by establishing images and video of key nematode pests of sugarcane. Then, a nematode separation kit will be designed based on the principles that ‘nematodes in a water suspension will settle at bottom due to specific gravity’. Finally, the AI algorithm developed will be embedded on a microscope set up designed with display unit. The kit proposed for nematode extraction from soil samples will make the separation process easy. Then microscopic image analysis is the core architecture of this project in which expert labeled huge image data sets will be used to generate AI-deep learning system. The implementation of this project will yield a novel nematode detection kit that ultimately make the sugarcane nematode diagnosis and monitoring as easy, faster and accurate process. |