Machine vision techniques for solar power forecasting and generation

该项目旨在利用机器视觉技术推进短期太阳能发电预测和优化发电过程的研究。该项目将使用相机捕捉天空和定日镜镜面的图像。科学创新是探索太阳能发电预测的几何感知特征表示，以及建立定日镜镜面的三维模型以优化太阳能发电过程。研究结果为提高太阳能发电预测精度提供了一个工作原型，并为太阳能发电提供了一个三维重建系统。这些成果对短期太阳能发电预测、发电和电网管理系统具有重要意义。

This project aims to advance the research in short-term solar power forecasting and optimise the generation process using machine vision techniques. This project will use cameras to capture images of sky and mirror surfaces of heliostats. The scientific novelties are the exploration of geometry-aware feature representations for solar power prediction and building three-dimensional models of mirror surfaces of heliostats to optimise the solar power generation process. The outcome is a working prototype to boost the solar power forecasting accuracy and a three-dimensional reconstruction system to be helpful for the solar power generation. These outcomes will highly benefit the short-term solar power forecasting, generation and electricity grid management systems.