9:00 AM - 9:15 AM
[HDS08-01] Discussion on the performance of consumer-grade LiDAR in landslides monitoring
Keywords:landslides, consumer-grade LiDAR, monitoring
In Taiwan, landslides happened frequently. The ability to anticipate signs of a collapse in advance could significantly reduce the financial and property losses caused by slope disasters. Therefore, monitoring is a crucial task for landslides. This study explores the potential of LiDAR technology for real-time monitoring. A consumer-grade LiDAR, the Livox Mid-70, is used in this experiment. Its accuracy is less than LiDAR valued in the millions, but is cost-effective and easy to install, making it suitable for established a long-term monitoring system in the field. To incorporate consumer-grade LiDAR into monitoring applications, this study seeks to investigates its performance and potential measurement errors during monitoring.
The price of Livox Mid-70 is approximately US$1000. Considering various measurement errors and experimental site limitations, we focuses on investigating the impact of point cloud density and temperature on measurements. In the experiment examining the impact of point cloud density, 1m x 1m cement board was placed at distances of 10m, 20m, 30m, and 40m. Scans were performed for 30 seconds to 5 minutes at each distance. Additionally, short-distance displacements of 50m, 70m, and 90m were tested by moving the cement board forward and backward by 1cm and 5cm. Due to battery limitations, the test was conducted in two parts, and although there were slight differences in environmental conditions, the analysis results still effectively demonstrated the impact of point cloud density.
For the temperature experiment, due to the inability to prepare an experiment space with completely controlled environmental temperature and a sufficiently long distance, the experiment was divided into two parts: testing the influence of instrument temperature and target temperature. In the instrument temperature experiment, a heat gun was used to heat the LiDAR from 20°C to 50°C, and we recorded data in 5°C increments. In the target temperature test, a 20cm x 20cm concrete board was heated to 50°C in a laboratory oven, and then, with natural cooling, the point cloud data for the concrete board was tested for signs of deviation.
In addition, this study monitored a building in Nantou Lushan that was expected to collapse for a month, aiming to understand whether this monitoring method could be applied to a moving surface. Scans were conducted on the building surface every night at 10 p.m. for one minute. From the point cloud results, it was observed that due to installation issues, the instrument rotated every day, and there were instances of the instrument being touched, causing displacement. Therefore, ICP was used to correct the position of the point cloud, but the results for some days were not satisfactory due to the relationship between point cloud intensity and density. After analyzing and excluding low quality data, it was concluded that there was no movement on the wall during these 30 days, and this result matched the GPS positioning.
Conclusions from this study are as follows:
1. Point cloud density is directly proportional to scanning time, increasing as scanning time extends.
2. There exists a linear relationship between point cloud density and standard deviation.
3. Target temperature does influence cement board measurements, resulting in a noticeable inward deformation.
4. Apparatus temperature does influence wall measurements, causing the wall to move backward as the temperature rises.
The price of Livox Mid-70 is approximately US$1000. Considering various measurement errors and experimental site limitations, we focuses on investigating the impact of point cloud density and temperature on measurements. In the experiment examining the impact of point cloud density, 1m x 1m cement board was placed at distances of 10m, 20m, 30m, and 40m. Scans were performed for 30 seconds to 5 minutes at each distance. Additionally, short-distance displacements of 50m, 70m, and 90m were tested by moving the cement board forward and backward by 1cm and 5cm. Due to battery limitations, the test was conducted in two parts, and although there were slight differences in environmental conditions, the analysis results still effectively demonstrated the impact of point cloud density.
For the temperature experiment, due to the inability to prepare an experiment space with completely controlled environmental temperature and a sufficiently long distance, the experiment was divided into two parts: testing the influence of instrument temperature and target temperature. In the instrument temperature experiment, a heat gun was used to heat the LiDAR from 20°C to 50°C, and we recorded data in 5°C increments. In the target temperature test, a 20cm x 20cm concrete board was heated to 50°C in a laboratory oven, and then, with natural cooling, the point cloud data for the concrete board was tested for signs of deviation.
In addition, this study monitored a building in Nantou Lushan that was expected to collapse for a month, aiming to understand whether this monitoring method could be applied to a moving surface. Scans were conducted on the building surface every night at 10 p.m. for one minute. From the point cloud results, it was observed that due to installation issues, the instrument rotated every day, and there were instances of the instrument being touched, causing displacement. Therefore, ICP was used to correct the position of the point cloud, but the results for some days were not satisfactory due to the relationship between point cloud intensity and density. After analyzing and excluding low quality data, it was concluded that there was no movement on the wall during these 30 days, and this result matched the GPS positioning.
Conclusions from this study are as follows:
1. Point cloud density is directly proportional to scanning time, increasing as scanning time extends.
2. There exists a linear relationship between point cloud density and standard deviation.
3. Target temperature does influence cement board measurements, resulting in a noticeable inward deformation.
4. Apparatus temperature does influence wall measurements, causing the wall to move backward as the temperature rises.