Satellite Imagery, AI identify major crops for hill areas
Source: Chronicle News Service / Premchand Thongam

Imphal, April 01 2025: As part of a pilot project jointly initiated by the Department of Horticulture in association with Manipur Remote Sensing Application Centre (MARSAC), major crops suitable to five hill districts of the state have been identified through satellite imagery.

Talking to this daily, Directorate of Horticulture and Soil Conservation director K Debadutta Sharma disclosed that the Department of Horticulture, in collaboration with the MARSAC, undertook a pilot project to identify major crops suitable for cultivation in Chandel, Churachandpur, Noney, Pherzawl and Tengnoupal districts, using satellite imagery based on the climatic conditions of the areas.

Directorate of Horticulture and Soil Conservation director K Debadutta Sharma

The project spanned over a year, and the final results have now been obtained.

According to the satellite imagery conducted by MARSAC, citrus fruits are viable in Chandel, pineapple in Churachandpur, banana in Noney, ginger in Pherzawl and 'Heiribop' in Tengnoupal.

Plantation of these crops are in process.

The director stated that the pilot project was initiated due to the necessity of collecting soil samples and climatic data from various parts of the districts.

However, the process involves navigating difficult terrains, posing significant challenges for scientists conducting the survey.

He further said that collecting soil samples entails extensive coverage across the hill districts rather than from a single location.

The samples must be gathered randomly from a vast area, adding to the complexity of the study.

Additionally, the survey demands a distinct methodology to ensure accurate data collection, necessitating more time and resources, said Debadutta.

Explaining how satellite imagery can help identify major crops suitable for the five hill districts, the director stated that data on the characteristics of various crops available in the state was fed into an artificial intelligence (AI) tool to determine suitable crops for different areas.

He further explained that the AI tool requires characteristic data of crops to generate accurate results.

In case of more data inputted, then the finding turns out to be comprehensive.

Once the data is fed into the system, it undergoes machine learning to enhance its predictive capabilities.

Initially, only a limited dataset for a few crops was provided to launch the pilot project.

As more data on other crops becomes available, the AI system will be able to identify the most suitable cultivation areas across the state, maintained the director.

However, Debadutta noted that the project is highly expensive.

In this context, experts explain that crop identification is being conducted using various index analyses, including satellite im agery spectral signatures and the Normalised Difference Vegetation Index (NDVI).

Different crops reflect light at varying wavelengths, creating unique spectral signatures that enable the identification of specific crops.

(This article is published as part of the 6th Scientific Journalism Programme, initiated by the Public Relations and Media Management (PRMM) division of Central Agricultural University, Imphal, under the theme "Artificial Intelligence and Its Uses") .