IIT Mandi develops AI based model for disease detection in potato crops using photos of leaves

Researchers on the Indian Institute of Technology (IIT) Mandi, have developed a computational mannequin primarily based on Synthetic Intelligence (AI) for automated illness detection in potato crops utilizing images of its leaves. The analysis in collaboration with the Central Potato Research Institute, Shimla, makes use of AI strategies to spotlight the diseased parts of the leaf and has additionally been printed within the journal — Plant Phenomics.

The computational software developed by IIT Mandi scientists can detect blight in potato leaf photos. The mannequin is constructed utilizing an AI software known as masks region-based convolutional neural community structure and might precisely spotlight the diseased parts of the leaf amid a posh background of plant and soil matter.

The researchers are additional engaged on changing the developed software to a smartphone software for a extra sensible utilization.

In line with the staff, potatoes, within the historical past of the world, have been the reason for the nice famine of the mid-nineteenth century that killed over 1,000,000 folks in Eire and rang the dying knell for the Irish language. The rationale? Potato Blight.

“The blight is a standard illness of the potato plant, that begins as uneven mild inexperienced lesions close to the tip and the margins of the leaf after which spreads into giant brown to purplish-black necrotic patches that finally results in rotting of the plant. If left undetected and unchecked, blight may destroy your complete crop inside per week beneath conducive circumstances,” mentioned Srikant Srinivasan, Affiliate Professor, College of Computing and Electrical Engineering, IIT Mandi.

“In India, as with most growing nations, the detection and identification of blight are carried out manually by skilled personnel who scout the sphere and visually examine potato foliage,” he mentioned including, this course of, as anticipated, is tedious and infrequently impractical, particularly for distant areas, as a result of it requires the experience of a horticultural specialist who is probably not bodily accessible.

Joe Johnson, analysis scholar at IIT Mandi defined that the automated illness detection might help on this regard and given the intensive proliferation of the cellphones throughout the nation, the smartphone may very well be a great tool on this regard.

“The superior HD cameras, higher computing energy and communication avenues provided by smartphones supply a promising platform for automated illness detection in crops, which may save time and assist in the well timed administration of ailments, in circumstances of outbreaks,” he mentioned.

For the analysis, so as to develop a strong mannequin, wholesome and diseased leaf information have been collected from fields throughout Punjab, Uttar Pradesh and Himachal Pradesh.

“It was vital that the mannequin developed ought to have portability throughout the nation. Evaluation of the detection efficiency signifies an total precision of 98 per cent on leaf photos in area environments,” Srinivasan mentioned.

The seven-member staff claimed despite the fact that potato will not be a staple meals in most areas of the world, it’s a money crop, and failure in it may have disastrous penalties, notably to farmers with marginal landholding.

Thus, early detection of blight is vital to stop monetary disaster to the farmer and the nation’s economic system.

“Following this success, we are actually sizing down the mannequin to a couple tens of megabytes in order that it may be hosted on a smartphone as an software. With this, when the farmer will {photograph} the leaf which seems unhealthy, the appliance will verify in real-time if the leaf is contaminated or not,” mentioned Srinivasan.

“With this well timed data, the farmer would know precisely when to spray the sphere, saving his produce and minimising prices related to pointless use of fungicides. The mannequin is being refined as extra states are coated,” added Srinivasan.

He additionally highlighted that it might be deployed as a part of the FarmerZone app that might be obtainable to potato farmers free of charge.

The opposite members of the staff embody Shyam K Masakapalli from IIT Mandi together with analysis students — Joe Johnson and Geetanjali Sharma, and Vijay Kumar Dua, Sanjeev Sharma and Jagdev Sharma from the Central Potato Analysis Institute, Shimla.

FbTwitterLinkedin