Artificial intelligence helping in the early detection of pests and diseases in crops

Vadodara: In what could be a major step towards food security by stopping crop damage, artificial intelligence (AI) and machine learning (ML) techniques are being used to gather knowledge about pests and diseases causing crop loss in cereals such as rice , corn, wheat and sorghum are useless. Thanks to a scheme funded by the Government of Gujarat, the Faculty of Basic Sciences and Humanities, Anand Agricultural University (AAU) has started research work in this direction. About 30% of the country’s annual crops are wasted due to pests. Some reports from the Indian Council of Agricultural Research say that the pests, along with nematodes, cause losses of 60 million tonnes of crops annually. The aim is to study the impact of diseases and insect pests on grain quality and diagnose such problems using various computer vision models. Data for the project will be collected from major research centers including the Central Rice Research Center at Navagam; Major Maize Research Center at Godhra; Agricultural Research Station at Arna (for wheat) and Central Research Station for Sorghum at Surat. In addition to the research centers, scientists will also collect information and images of crop pests and diseases from farmers’ fields located near the centers. “The results of this research will enable farmers to accurately identify diseases and insect pests attacking these vital cereal crops and lend a hand in their earlier protection. Farmers will be trained and instructed in this,” said AAU vice-chancellor Dr KB Kathiria, who is spearheading the project along with director research Dr MK Jhala and director and dean Dr YM Shukla. As part of the project, image data will be prepared by taking detailed photos of pests and diseases that cause damage depending on the seasons and different stages of cultivation. These photos will then be processed. The resulting images will then be used as input to machine learning (ML) methods such as neural networks and deep learning models, and the model will be trained and analyzed. “Images of crop pests and diseases will be generated using large-configuration cameras and an image database will be generated,” he said. Data processing will be carried out using, among others, a support vector machine, an artificial neural network, the nearest neural network, a probabilistic neural network and a back-propagation neural network. Each model is expected to report its results. “These models have an accuracy ranging from 75 to 99.99%, which means the pest diagnosis using this model is between 75 and 99.99%,” the official said.