IIT Delhi develops landslide extent mapping tool based on cloud computing, machine learning

NEW DELHI: Researchers from the civil engineering department at the Indian Institute of Technology (IIT) Delhi have developed a cloud computing and machine learning-based tool, ML-CASCADE, to map landslide extent using satellite data.

The user-friendly tool available in the public domain requires only two inputs – an approximate date and location of a landslide event. It can accurately map a complex cluster of landslides within five minutes and a simple landslide within two minutes, which is crucial for post-disaster damage assessment. The underlying model is trained on a diverse set of data including a large amount of satellite, terrain, vegetation, and soil characteristics.

A research paper on this innovation, authored by PhD scholar Nirdesh Kumar Sharma and professor Manabendra Saharia from the HydroSense Lab at the civil engineering department, has been published in the ‘Landslides’ journal by the International Consortium on Landslides (ICL).

“Traditionally, landslides have been mapped by manually digitizing over satellite imagery, which is costly, inaccurate, and time-consuming. Field surveys and geological data collection cannot be done in large and remote areas. Existing simple models developed using the thresholds of vegetation indices fail in areas with minimal vegetation. Machine learning on geospatial data offers an unprecedented opportunity to overcome the drawbacks of index-based methods and integrate multiple diverse datasets to map landslides with high accuracy”, said professor Saharia.

Also read BHU scientists secure patent for Ayurvedic formulation inhibiting over 95% Covid virus growth

Landslide inventory

A key innovation of ML-CASCADE is its dynamic nature; it builds custom models tailored to user inputs and local environmental factors rather than relying on pre-trained models, which often lack sufficient annotated training data for landslide applications. This adaptability allows for accurate modeling across diverse terrains.

Designed with input from both technical and non-technical users, the tool features an intuitive interface. Outputs can be easily downloaded in multiple formats for use in further GIS analysis. By harnessing Google Earth Engine as its backend, ML-CASCADE avoids the need for data downloads, utilizing parallel processing to generate results in minutes—significantly faster than conventional local computing methods.

The tool has been rigorously validated over thousands of landslides, with two major events in Himalayas (Kotrupi landslide, 2017) and Western Ghats (Kodagu landslide, 2018) provided as case studies in the paper. Many landslides go unmapped, especially in remote areas, and developing a comprehensive landslide inventory typically requires extensive expertise and resources. IIT Delhi researchers are employing this tool to create a national historical landslide inventory, which will be essential for developing early warning systems.

Looking ahead, the innovative architecture of ML-CASCADE holds promise for broader applications, including mapping flood inundation, deforestation, sand mining, and other pressing environmental challenges.