Telangana government scouts for cloud-based AI solution to track progress of Indiramma Illu

State government seeks AI solution to track housing allotment and construction progress for Indiramma Illu programme efficiently.

The State government is scouting for a comprehensive artificial intelligence based solution that can help officials track allotment of new houses under the first phase of the Indiramma Illu, weaker section housing programme, and also monitor the progress of house construction.

The government is looking at a cloud-based AI construction detection model which would be able to take input from various sources such as housing app inside smart phone or a camera live feed. The proposal has been mooted as the programme involves lakhs of beneficiaries whose identification can be made by looking at their current dwelling status.

The development follows the decision to construct 3,500 houses each in the 119 Assembly constituencies totalling 4.5 lakh houses in the first phase for beneficiaries who own a piece of land. The government has announced financial assistance of ₹5 lakh in four instalments for construction of each house to eligible beneficiaries and budgetary provision of ₹22,500 crore had been made towards Indiramma Illu.

Deployment of field surveyors across the State is under active consideration so that they will use the application integrated into the AI model for verifying each beneficiary. The system collects all the data and assignment of resources will be dependent on various factors which the Housing department will decide later. Once the allocation is done, the software application should track the construction progress based on which monies would be released to various contractors and agencies.

The proposed cloud-based solution makes it mandatory that the image should be from the same address where the survey of the construction is being done besides validating the geo-coordinates in the photograph. The images/feeds sent to the solution should confirm the image type whether the house is under construction or fully constructed besides confirming the status of the unit as dwelling or non-dwelling.

The images sent to the solution should confirm one of the four stages the model has been trained in and there should be a dashboard that would inform authorities concerned where progress had not been made and the model should support at least 15,000 hits at any given point of time. The solution should be flexible to be trained with new stages or to improve accuracy.

The IT department is learnt to have issued a request for proposal inviting companies for the deployment, implementation, integration and support of a cloud-based model to validate different stages of house construction enabling release of payments to vendors.