Summary

• Our client is from the solar industry who are majorly looking for a solution for their solar panel services. Their solution uses advanced big data techniques, image and video processing through usage of drones.
• The client was clear on their requirements about real-time image and video processing results and the solution has been built to offer real-time capabilities using computer vision.
• Our team designed and implemented the solution using five plus technologies which included TensorFlow, TensorRT, Gstream, Python, Computer Vision and advanced Aritificial Intelligence techniques.

Tech Challenge

• A key challenge of the solution was to give interactive and real time reports of the malfunction.
• Another challenge of the solution was to map the solar panel to its correct position according to the map.
• The solution also required binary segmentation of panels through video processing.

Solution

• The solution is capable enough to track and segment the data using advanced video analytics.
• It can easily process and optimize the data with map coordinates using Big Data technologies and TensorRT.
• With use of computer vision, the solution is capable to give timely detection of any defect in the solar panels.

Project Planning

• 2 Week -Solution Architecture Desgin -Solution Architect

• 4 Week - Integration, Testing and Deployment

• 6 Week - Exploratory Data Analysis and Model Building

• 8 Week - Collection of data and realtime streamlining and development

Results

• We have developed a system that allows one time accurate routine checkup of the solar panels.
• With the use of computer vision and advanced big data our customers are now able to get detailed analysis reports of the inspection.
• With technologies like TensorFlow customers can now get a summary highlighting potential defects, energy and money losses.