PRC STAR 41603SA
Project lead from ImViA : Albert Dipanda, Cyrille Migniot Partners :Professeur Gwanggil Jeon from EISP (Embedded Image & Signal Processing) of university of Incheon. Beginning : 01/06/2018 End : 31/05/2020
In this project, we propose an extensive study of the point cloud denoising to define new algorithms that provide 3D high precision representation of a scene. The data is acquired by a multi-kinect system which allows to obtain a complete point cloud. In the Le2i laboratory, we have developed a new method for human detection from a complete point cloud. We have proposed a descriptor adapted for point clouds: an histogram defines the spatial repartition of the points according to a cylindrical representation. Although the algorithm has demonstrated its reliability and a high accuracy, noise on the acquisition could define a weak point. First, depth map gives holes (part that could not been computed) that limit the amount of information about the scene. Secondly some aberrant points could appear and decrease the descriptor representation. Finally the front view of the sensor provides an unequal spatial repartition of the points (parts that are parallel to the view axis are represented by less points than the parts perpendicular to this axis). These drawbacks could affect the description of the histogram and then reduce the efficiency of the method. Our goal is to remove these three artefacts and to check the improvement of the scene analysis method. In the EISP laboratory, we have developed demosaicking and denoising methods for color images. Using our experience in denoising
Abstract :
methods, we will develop denoising technique in point cloud.
- kc_data:
- a:8:{i:0;s:0:"";s:4:"mode";s:2:"kc";s:3:"css";s:0:"";s:9:"max_width";s:0:"";s:7:"classes";s:0:"";s:9:"thumbnail";s:0:"";s:9:"collapsed";s:0:"";s:9:"optimized";s:0:"";}
- kc_raw_content:
- extrait:
- lien_externe:
- equipe:
- a:1:{i:0;s:5:"CORES";}
- tags:
- PRC
- nom_du_projet:
- PRC STAR 41603SA : Denoising Technique in Point Cloud
- porteur:
- https://imvia.u-bourgogne.fr
- date_de_debut:
- 02/05/2020