An UAV based Autonomous Early warning System for Landslide The technical community concerned to disaster preparedness is facing challenges in arriving at opportune decisions requiring huge quantities of quality geo-spatial data, in real time and/or near RT. Unmanned Aerial Vehicle (UAV) technology, having its advantages of rapid deployability, mobility and low cost, is emerging out as a potential solution. In this study, a UAV based Automated System, capable of warning for an impending landslide, is being proposed.The system under development involves platform and payload design and operation, project planning, post-processing and disaster application of the acquired data.

The development of proposed system consists of two broad divisions: the hardware and the software. Hardware is a UAV based geo-spatial data acquisition system consists of sub-systems such as UAV platform, integrated GPS/INS system, data acquisition sensors and wire less communication device etc. Software developemnt involves two aspects: data pre-processing, to make available quality data, and data application, to arrive at required decision using the aquired pre-processed data. The data collected by the UAV sensors has been planned to be communicated to the field office through wirless communication. The pre-processing software will carry out data visualisation and quality checking/control, geo-referencing, orthorectification, radiometric corrections, bundle adjustment, mosaic preparation, DTM extraction, DSM preparation etc. The application software will consist of machine learning based module for disaster analysis and relevant decision process. Further, a real-time processor installed in UAV will also proclaim warning messages through wireless.

The challenges those are/expected to facing during development of the system are integration and synchronization of different sensors, stabilization and control of UAV system, determination of the position and orientation offset between the GPS/INS system and camera, clock synchronization of GPS/INS system with camera, offset and orientation during each mission, interfacing of modules, quality of geo-data, wireless data transmission and receiving, radiometric variations in differrent frames of geo-data, data handling formats, real-time/near real-time management and analysis of huge volume of acquired data etc.

The modular structure of the proposed system will be extended further to take care of the monitoring, mangement and rescue as well as rehabilitation during landslide in real-time. Further, the system may be extended for other disasters such as flood, forest fire etc simply by incorporating a specific module for each type of disaster.