UAV analytics in support of geothermal energy exploration

该研究项目通过开发分析工具来支持地热能勘探 将无人机(UAV)获取的红外图像转化为信息产品。无人机是 有望取代通常用于识别和绘制地表热量异常的其他遥感平台 在地热勘探中，特别是对于占地面积高达几平方公里的项目，因为它们的面积较低 制作高分辨率地热热点地图的成本、可操作性和能力。然而，在那里 是将无人机红外图像融入地热勘探的重大技术障碍 工作流程，这是卡尔加里大学和北极星大学之间合作的动力 GeopPower-加拿大领先的地热勘探公司之一。解决这个问题的主要挑战是 研究是适应和定制计算机视觉算法，以便无缝集成多达几个 由无人机获取的数百幅红外图像转换成几何和辐射精确的地图 地表热量异常。为载人红外图像采集开发的常规分析工具 飞机和卫星并不直接适用于无人机，因为复杂的几何和 这些低空飞行平台获取的图像中的辐射失真。 NSERC Engage基金将用于开发一个分三个阶段的研究项目：(1)算法开发； (2)盲法实验；(3)半自动加工工作流。第一阶段将涉及测试和 对计算机视觉研究中的特征匹配和图像归一化算法进行定制。 第二阶段将在瓦勒蒙特附近的独木舟里奇地热项目进行盲目实验， 公元前。阶段3将算法或算法集转换为定制软件工具。总体而言，主要 该项目的贡献是开发了新的分析方法，扩展了用于识别、 绘制和评估地热能资源图。

This research project supports geothermal energy exploration through the development of analytical tools that translate infrared images acquired by an unmanned aerial vehicle (UAV) into information products. UAVs are expected to replace other remote sensing platforms commonly used to identify and map surface heat anomalies in geothermal exploration, particularly for projects with footprints up to several square-km, due to their lower cost, maneuverability, and capacity to produce high-resolution maps of geothermal hotspots. However, there are significant technical barriers to the integration of UAV infrared imagery into geothermal exploration workflows, which is what motivated this collaboration between the University of Calgary and Borealis GeoPower - one of Canada's leading geothermal exploration companies. The main challenge tackled in this research is to adapt and customize computer vision algorithms in order to seamlessly integrate up to several hundred infrared images acquired by a UAV into a geometrically-accurate and radiometrically-precise map of surface heat anomalies. Conventional analytical tools developed for infrared imagery acquired by manned aircraft and satellites do not straightforwardly apply to UAVs because of the complex geometric and radiometric distortions in imagery acquired by these low-flying platforms. NSERC Engage funding will be used to develop a three-stage research project: (1) Algorithm development; (2) Blind experiment; and (3) Semi-automated processing workflow. Stage 1 will involve testing and customization to feature matching and image normalization algorithms adapted from computer vision research. Stage 2 will test the analytics in a blind experiment at the Canoe Reach Geothermal Project near Valemount, BC. Stage 3 will translate the algorithm or set of algorithms into custom software tools. Overall, the main contribution of this project is the development of new analytics that extend the breadth of tools used to identify, map and assess geothermal energy resources.