Optical Navigation for Heliostat

Abstract

The master thesis deals with development a low cost optical image sensor System which can be used as position sensor for controlling the Heliostat/Solar mirrors. Heliostat is an object which reflects light of Sun to the central tower, reflected light from heliostats is then further used to generate electricity. An optical image sensor with an associated microcontroller/computer determines the position of the Heliostat relative to Sun and central tower and calculates the necessary tracking movement. Prototype accuracy of � 0.6 - 3.2 mrad, costs � 100  /optical sensor for large quantity and acceptance angle � 60° is achieved. The work package can be distributed in five phases. In the initial phase of the project with the help of simulation result from MATLAB, specifications of the sensor is demonstrated. Initial phase also includes vigorous search of the available Cameras and comparison between them. Selection of micro-controller is also part of this phase. In the second stage camera is calibrated with two different algorithms and results are presented. In the third phase more focus is placed towards handling of all the hardware in same environment. For this reason visual studio 2013 C++ is chosen. This phase includes task such as image acquisition, pre-processing and setup of serial communication protocol for Arduino. Fourth stage is dedicated for image processing. In this stage task such as Sun detection, marker detection and pose estimation are represented. This stage also generates necessary time frame for the motor movement according to the offset. Fifth stage is dedicated to the testing and validation of the result. In the last stage future possibilities of the project are presented.