Examinando por Materia "LiDAR"
Mostrando1 - 4 de 4
Resultados por página
Opciones de clasificación
- PublicaciónSólo datosA Comparative Study of 3D Plant Modeling Systems Based on Low-Cost 2D LiDAR and Kinect(Lecture Notes in Computer Science, 2021-05-16) Murcia, Harold; Sanabria, David; Méndez, Dehyro; Forero, Manuel G.Morphological information of plants is an essential resource for different agricultural machine vision applications, which can be obtained from 3D models through reconstruction algorithms. Three dimensional modeling of a plant is an XYZ spatial representation used to determine its physical parameters from, for example, a point cloud. Currently two low-cost methods have gained popularity in terms of 3D object reconstructions in 360 ∘ employing rotating platforms, based on 2D LiDAR and Kinect. In this paper, these two techniques are compared by getting a 3D model of a Dracaena braunii specie and evaluating their performance. The results are shown in terms of their accuracy and time consumption using a Kinect V1 and a LiDAR URG-04LX-UG01, a well-performance low-cost scanning rangefinder from Hokuyo manufacturer. In terms of error calculation, the Kinect-based system probed to be more accurate than the LiDAR-based, with an error less than 20% in all plant measurements. In addition, the point cloud density reached with Kinect was approximately four times higher than with LiDAR. But, acquisition and processing time was about twice than LiDAR system.
- PublicaciónSólo datosComparative study of point cloud registration techniques between ICP and others(Proceedings of SPIE - The International Society for Optical Engineering, 2020-08-21) Méndez, Dehyro; Forero, Manuel G.; Murcía, Harold F.Registration is a technique employed for the alignment of point clouds in a single coordinate system. This process is very useful for the reconstruction of 3D plant models, the extraction of their morphological features and the subsequent analysis of the phenotype. One of the most widely studied recording algorithms is ICP (Iterative Closest Point), which is based on rigid transformations. Although in the literature there are several comparative studies between different variants of ICP, there is no comparative study with other more recent existing methods based on other principles. Therefore, in this paper we present a study comparing the results obtained with different registration algorithms on previously filtered 3D point clouds of plants, obtained with a MS Kinect V1 sensor integrated to a rotating base. The study includes two of the most used variants of the ICP, the point-to-point ICP and the point-to-plane ICP. These variants are based on the normals to the surfaces found to guide their point-to-point matching method presenting better results in smooth regions. In addition, other iterative point cloud alignment algorithms based on probability density estimation, hierarchical mixed Gaussian models and distance minimization between probability distributions are included. The results showed the effectiveness of ICP variants simplicity, and the high precision achieved by probabilistic methods. The error and computation time of the algorithms, implemented in Python, were evaluated.
- PublicaciónSólo datosDevelopment of a Simulation Tool for 3D Plant Modeling based on 2D LiDAR Sensor(2020 Virtual Symposium in Plant Omics Sciences, OMICAS 2020, 2021-09-12) Murcia, Harold F.; Sanabria, David A.; Méndez, DehyroThe three-dimensional modeling of plants allows not only the use of color information, as in conventional digital image processing, but also the use of geometric information for the morphological extraction of their features and the subsequent analysis of their phenotype. The generation of point clouds is one of the initial stages of this process, which is carried out in different ways. One of the techniques used for this purpose uses a rotating platform and laser sensors, which employ multiple beams of light to illuminate the measurement area and determine its depth with the principle of time of flight (ToF). However, the algorithms used to perform the three-dimensional reconstruction must be calibrated in a process that may include a large number of experiments. For this reason, artificial three-dimensional point clouds generated by simulators may be suitable, both for the validation of reconstruction algorithms on those platforms and for the analysis of plant phenotype characteristics under almost realistic conditions. Thus, with this aim, this paper describes the development of an open-source tool for the generation of artificial 3D plant point clouds, based on the simulation tool Gazebo and the Robot Operating System (ROS). This work in progress allows validating different reconstruction algorithms, as well as the characteristics of LiDAR sensors and turntables to generate 3D models in an open file format. Our source implementation is freely available online and can be obtained from https://github.com/HaroldMurcia/3D-plantModeling-with-2DLiDAR.
- PublicaciónAcceso abiertoReconocimiento de elementos de escena sobre nubes de puntos tridimensionales en aplicaciones de robótica móvil terrestre.(Universidad de Ibagué, 2019) Florez Romero, Andres Mauricio; Murcia Moreno, Harold FabiánThis project is developed in the seedbed SI2C of the research group D+TEC of the Universidad de Ibagué, it consists on the a supervised classification algorithm for scene elements recognition on urban environments, taking points clouds acquired with an acquisition system that implements in LAAS (Laboratory for Analysis and Architecture of Systems) that implements a LiDAR sensor on board a UGV. The classifiers were Random Forest (RF), Support Vector Machine (SVM), AdaBoost, Neural Network (NN) and Gaussian Naive Bayes, those were trained with 3D features, 2D and characteristics of full wave form information, the last one being the contribution generated for projects already developed. The classifiers were evaluated using the F1 Score indicator, obtaining an average of scores up to 86% accuracy getting a classifier with rates of high success and demonstrating the contribution provided by the descriptors proposed based on the information obtained from the full wave characteristics, respect to conventional descriptors