-Recently it is observed that there is an increasing trend of using lidar point clouds in construction, since lidar is able to provide highly accurate 3-dimensional representations of objects. The flip side of a lidar point cloud is its massive size and therefore prolonged time for data processing. Our research aims to develop an efficient framework for as-built modelling in terms of time, cost and performance. For this, we collected lidar point clouds using two different data collection methods (i.e. mobile and terrestrial lidar systems with a few centimetre level accuracy) and constructed solid 3-dimensional models of a building. The procedures of creating asbuilt models from both data acquisition methods are compared to understand the capability of each for automatic building information modelling. In order to create as-built models, a framework consisting of eight key stages from data capturing to constructing the building information model is developed. It was found that the framework using mobile lidar enables contractors to create as-built models for complex objects in a timely manner, whereas the framework using terrestrial lidar provides us with more accurate as-built information models. The implementation results of the two frameworks using mobile and terrestrial lidar systems vary between 5-30 mm and 1-45 mm, respectively. It is anticipated that the proposed study becomes a step forward to full automation of lidar-based building information modelling.

No abstract

Forest fires are representative natural disasters resulting from dramatic global climate change in these modern times. When forest formation is insufficient due to forest damage caused by fire, secondary damages such as landslides occur during the winter thawing period and heavy rains. In most countries, only a limited area is managed as CCTV-centered monitoring systems for forest management. For the landslide prediction, markers containing 3D spatial coordinates were located on the slopes of the danger areas in advance. Then 3D mapping and angle of repose were obtained by periodic drone imaging. The recognition range and angle of view of markers were defined, and a new method for predicting signs of landslides in advance was presented in this study.