Mines Repository
Communities in Mines Repository
Select a community to browse its collections.
Recent Submissions
Publication Open Access Mines library: a newsletter for friends and supporters of the Arthur Lakes Library, Fall 2025 - Winter 2026(Colorado School of Mines. Arthur Lakes Library, 2025-12)Publication Open Access An algorithm for rockfall block reconstruction from point clouds of debris fragments(Colorado School of Mines. Arthur Lakes Library, 2026-04)As rockfalls travel down a slope, they often break into fragments. Determining how fragments originally fit together within their source block has the potential to lead to new insights regarding the fragmentation process, and help to constrain numerical fragmentation models. While digital item reconstruction from fragments is conducted in archaeology and medicine, according to our review of the literature, similar approaches have not been applied to rockfall. In this study, we develop a method to reconstruct rockfall blocks from their fragments and test it using a synthetic case. We simulated fragments by taking lidar scans of two cardboard shapes and a compound shape made from those two blocks. The scans were combined into one point cloud and each fragment’s point cloud was used to make a watertight shape. We developed and tested a process to automatically manipulate fragment positions and optimize their arrangements to minimize overlap between blocks and protrusion from the source block. Ultimately we were unable to verify that digital reconstruction is feasible with this procedural approach and optimization function. However, the investigation highlighted multiple areas for further investigation: error metric improvements and optimization algorithm improvements.Publication Open Access SMR-GC: matching a dashcam against satellite imagery for GPS-denied localization(Colorado School of Mines. Arthur Lakes Library, 2026-04)GPS-denied localization in urban environments is of vital importance. However, existing localization solutions are mostly dependent on stereo cameras, reference maps, and location calibration. In this context, we propose a novel visual localization framework called Spectral Multi-scale Registration with Geometric Constraints (SMR-GC) for GPS-denied navigation. Our proposed framework is capable of achieving sub-meter accuracy with only a monocular dashcam and automatically generated aerial reference maps. Our framework is composed of four main stages: (1) real-time semantic segmentation of road, building, and vegetation classes from dashcam images using SegFormer-B0, (2) inverse perspective mapping of perspective view to bird’s eye view, (3) spectral cross-correlation with Hessian uncertainty estimation of ground-level semantic maps with aerial reference maps, and (4) an SE(2) Extended Kalman Filter that combines vehicle odometry with registration measurements that are weighted by the curvature of the correlation surface. Our main contributions are: (1) we propose an automatic aerial reference map generation from satellite imagery without any manual labeling, (2) GPS and compass-based auto-calibration without any location calibration, and (3) proposed measurement-adaptive noise modeling of Extended Kalman Filter with noise modeled from registration correlation Hessian. We evaluate our proposed framework with five KITTI Odometry sequences with a rigorous protocol of full sequence evaluation with correlated drift odometry. Our preliminary results show that our proposed framework can achieve competitive accuracy with existing localization solutions using stereo cameras with only a single camera.Publication Open Access Evaluating thermal comfort of cross-laminated timber and lightweight construction during outage conditions(Colorado School of Mines. Arthur Lakes Library, 2026-04)Extreme weather events and power outages are happening more often in Colorado, which raises questions about how well buildings can keep indoor conditions comfortable without heating or cooling systems running. This research examines how construction materials affect indoor temperature stability during these events by comparing Cross-Laminated Timber (CLT) and lightweight (LW) construction. Using both building energy simulations and measured data, this study compares how each construction type performs during heating outages and heatwave conditions. The focus is on how quickly indoor temperatures change, how thermal mass helps slow down those changes, and the extent to which outdoor conditions affect indoor temperatures. The analysis also studies how quickly temperatures drop or rise and how long the indoor space stays outside a comfortable range. Preliminary results indicate that CLT tends to keep indoor temperatures more stable because of its higher thermal mass, meaning temperatures change more slowly than in lightweight construction. This can delay how quickly a space becomes uncomfortable during a power outage. This is especially important in high wind and dry conditions, where Public Safety Power Shutoffs (PSPS) can leave people without power for extended periods. Extreme weather is straining our existing grid infrastructure, which makes improving resiliency on the demand side important to protect against the risk of increasing grid outages in the future. This work connects material choice to real performance outcomes and helps inform more resilient and energy-efficient design.Publication Open Access Portfolio of photographs of the Down Town mines, Leadville, Colorado(Colorado School of Mines. Arthur Lakes Library, 1915-06)