Machine Learning Based Feature Extraction of Electrical Substations from Satellite Data Using Open-Source Tools


Machine learning comes under Artificial intelligence where a machine mimics the human brain in processing the data for various purposes like detecting objects, recognizing speech, translating languages, extracting features and making decisions. Feature extraction of the Power infrastructure using Machine learning techniques is a new challenge since much attempts have not been done in this field.

IEEE-ICETCI 2021 is organizing a Virtual Competition in association with RRSC-Central, NRSC, ISRO, Nagpur on ‘Machine learning based feature extraction of Electrical Substations from Satellite data using Open-Source tools’, and will take place from 1 March to 30 July, 2021.

The task of this competition is to develop a Machine learning-based software using open source tools. Further extract Electrical Substations from high resolution satellite data, and submit a paper (PDF) describing the techniques employed in solving the problem.

Evaluation will be done in two step process. In the first step, top 10 entries will be selected based on a metrics to evaluate correct identification and exact demarcation of substation boundary/extent. In the second step, for the top 10 entries, a code review will be done, software model efficiency will be tested followed by virtual presentation; the best three will be selected as winners of the competition.

The contestants will be provided with the Training Dataset of 100 Satellite data chips of ~1m resolution for training. Each image chip will have one electrical substation feature. Set of Points and Polygons AOI will also be provided for training the Machine learning network. A Testing Dataset of 20 satellite data chip’s mosaic, containing substations and other features will be provided for testing. The Tutorial Links will be provided for online learning of the API’s and libraries.

The competition data can be downloaded from here.

Competition Guidelines/Rules can be found here.

  Competition Flyer  


Anju Bajpai, T P Girish Kumar, Ashish Shrivastava, D S Prakasa Rao, Subrata N Das, G Sreenivasan, C S Jha
Regional Remote Sensing Centre-Central,
National Remote Sensing Centre, Amravati Road, Nagpur. /

Sangeeta Rajankar, Sapna Deotale, Dilip M Kolte
Maharashtra Remote Sensing Application Centre,
Government of Maharashtra, Nagpur.