Millions Of 3D Points Research Articles

Large-scale field phenotyping using backpack LiDAR and CropQuant-3D to measure structural variation in wheat.

Plant phenomics bridges the gap between traits of agricultural importance and genomic information. Limitations of current field-based phenotyping solutions include mobility, affordability, throughput, accuracy, scalability, and the ability to analyze big data collected. Here, we present a large-scale phenotyping solution that combines a commercial backpack Light Detection and Ranging (LiDAR) device and our analytic software, CropQuant-3D, which have been applied jointly to phenotype wheat (Triticum aestivum) and associated 3D trait analysis. The use of LiDAR can acquire millions of 3D points to represent spatial features of crops, and CropQuant-3D can extract meaningful traits from large, complex point clouds. In a case study examining the response of wheat varieties to three different levels of nitrogen fertilization in field experiments, the combined solution differentiated significant genotype and treatment effects on crop growth and structural variation in the canopy, with strong correlations with manual measurements. Hence, we demonstrate that this system could consistently perform 3D trait analysis at a larger scale and more quickly than heretofore possible and addresses challenges in mobility, throughput, and scalability. To ensure our work could reach non-expert users, we developed an open-source graphical user interface for CropQuant-3D. We, therefore, believe that the combined system is easy-to-use and could be used as a reliable research tool in multi-location phenotyping for both crop research and breeding. Furthermore, together with the fast maturity of LiDAR technologies, the system has the potential for further development in accuracy and affordability, contributing to the resolution of the phenotyping bottleneck and exploiting available genomic resources more effectively.

Unsupervised Clustering Method for Complexity Reduction of Terrestrial Lidar Data in Marshes

Accurate characterization of marsh elevation and landcover evolution is important for coastal management and conservation. This research proposes a novel unsupervised clustering method specifically developed for segmenting dense terrestrial laser scanning (TLS) data in coastal marsh environments. The framework implements unsupervised clustering with the well-known K-means algorithm by applying an optimization to determine the “k” clusters. The fundamental idea behind this novel framework is the application of multi-scale voxel representation of 3D space to create a set of features that characterizes the local complexity and geometry of the terrain. A combination of point- and voxel-generated features are utilized to segment 3D point clouds into homogenous groups in order to study surface changes and vegetation cover. Results suggest that the combination of point and voxel features represent the dataset well. The developed method compresses millions of 3D points representing the complex marsh environment into eight distinct clusters representing different landcover: tidal flat, mangrove, low marsh to high marsh, upland, and power lines. A quantitative assessment of the automated delineation of the tidal flat areas shows acceptable results considering the proposed method is unsupervised with no training data. Clustering results based on K-means are also compared to results based on the Self Organizing Map (SOM) clustering algorithm. Results demonstrate that the developed multi-scale voxelization approach and representative feature set are transferrable to other clustering algorithms, thereby providing an unsupervised framework for intelligent scene segmentation of TLS point cloud data in marshes.

City-Scale Localization for Cameras with Known Vertical Direction

We consider the problem of localizing a novel image in a large 3D model, given that the gravitational vector is known. In principle, this is just an instance of camera pose estimation, but the scale of the problem introduces some interesting challenges. Most importantly, it makes the correspondence problem very difficult so there will often be a significant number of outliers to handle. To tackle this problem, we use recent theoretical as well as technical advances. Many modern cameras and phones have gravitational sensors that allow us to reduce the search space. Further, there are new techniques to efficiently and reliably deal with extreme rates of outliers. We extend these methods to camera pose estimation by using accurate approximations and fast polynomial solvers. Experimental results are given demonstrating that it is possible to reliably estimate the camera pose despite cases with more than 99 percent outlier correspondences in city-scale models with several millions of 3D points.

Distinctive 2D and 3D features for automated large-scale scene analysis in urban areas

We propose a new methodology for large-scale urban 3D scene analysis in terms of automatically assigning 3D points the respective semantic labels. The methodology focuses on simplicity and reproducibility of the involved components as well as performance in terms of accuracy and computational efficiency. Exploiting a variety of low-level 2D and 3D geometric features, we further improve their distinctiveness by involving individual neighborhoods of optimal size. Due to the use of individual neighborhoods, the methodology is not tailored to a specific dataset, but in principle designed to process point clouds with a few millions of 3D points. Consequently, an extension has to be introduced for analyzing huge 3D point clouds with possibly billions of points for a whole city. For this purpose, we propose an extension which is based on an appropriate partitioning of the scene and thus allows a successive processing in a reasonable time without affecting the quality of the classification results. We demonstrate the performance of our methodology on two labeled benchmark datasets with respect to robustness, efficiency, and scalability.

Automatic registration of large range datasets with spin-images

Abstract Terrestrial laser scanning technique has represented one of the more advances occurred in the last years in the field of data acquisition. Time-of-Flight (TOF) systems provide a fast and reliable tool to measure millions of 3D points allowing a very effective and dense measurement of the surface geometry. Nowadays, the generation of high quality 3D models is a practice applied to different kind of objects: small or medium size artworks, parts of human body, cars, buildings, civil infrastructures (like dams, bridges, plants, etc.) and whole archaeological sites as well. In most cases, in order to capture the whole object geometry a number of single scans need to be acquired from different positions and then stitched together (i.e. registered each other) to generate the full 3D model. The automatization of the registration of multiple scans acquired from a terrestrial laser scanner (TLS) still represents a very attractive research field. The chance to automatically align several point clouds would reduce processing costs in terms of time and human resources. In addition it would allow even non-specialist users to produce 3D models with good quality. This paper contributes to this research area by presenting a method for the automatic registration of very dense point clouds acquired by TLS systems. The proposed solution is an extension to large datasets of an automatic range data registration procedure we developed a few years ago for the modelling of point clouds acquired with close-range laser scanners. Such procedure, based on the spin-images (SIs) algorithm, has been then improved with the introduction of a multi-resolution method that generates a pyramid of spin-images in order to speed up the matching between adjacent scans. The results we present show that this method can be successfully applied for the automatic registration of high density laser scans of complex and large structures of Cultural Heritage.