Very little research into the automated monitoring of unorganised stores has been developed to date. 80% of current warehouses are manually operated, but only 5% are totally automated. The work presented in this paper contributes with a new point of view regarding the store monitoring field related to the reconstruction of semantic 3D models of indoor scenes. Our system collects dense coloured points of the scene from different positions of a 3D laser scanner. Firstly, the accumulated point cloud is segmented into merchandise points and building-structure points using the MLSAC algorithm. The merchandise points are again segmented and classified according to a particular object model database. The system has been trained using a supervised learning algorithm with patterns based on the shape and colour of the merchandise object. The output is a 3D semantic model of the scene in which the recognised objects are placed in precise positions. An assessment of the storehouse in terms of stock, occupied/non-occupied volumes and a suggestion of the placement of new incoming packages is also carried out. The scanning platform is composed of a mobile robot carrying a 3D laser scanner and a DLSR camera that follows a pre-established path for each specific scenario. This approach has been tested with real data and evaluated against a ground-truth model in simulated scenarios, and has been shown to provide high recognition rates and low positioning errors.

Door detection is becoming an increasingly important subject in relation to autonomous mobile robot navigation in indoor environments. This paper presents an original approach that recognizes open and closed doors in 3D laser scanned data. The proposed technique uses both the geometric (i.e. XYZ coordinates) and colour (i.e. RGB/HSV) information provided by a calibrated set of 3D laser scanner and a colour camera. In other words, our technique is developed under a 6D-space framework. The geometry/colour integration and other characteristics of our method make it robust under occlusion and efficient to slight changes in door colours resulting from varying lighting conditions experienced from different scanning locations. The approach is tested in both simulated and real scenes, yielding encouraging results.

Object recognition may be a hard computer vision task under severe occlusion circumstances. This problem is efficiently solved in this paper through a new 3D recognition method for free-form objects. The technique uses the Depth Gradient Image Based on Silhouette representation (DGI-BS) and settles the problem of identification-pose under occlusion and noise requirements. DGI-BS synthesizes both surface and contour information of an object avoiding restrictions concerning the layout and visibility. Object recognition is carried out by means of a simple matching algorithm in the DGI-BS space which yields a point-to-point correspondence between scene and model. The method has been successfully tested in real scenes under occlusion and noise circumstances.

Manipulator robots working in changing environments need special path planners that provide good solutions in short times. In this work we describe a sample-based path planning algorithm that is suitable for such environments. Our approach is based on two points. Firstly, unlike most sampling techniques, we have defined a non-random local sampling which reduces the computational time greatly. Secondly, a set of interpolated walks is obtained through cubic splines which guarantee the smoothness and continuity of the walks. This path planning algorithm is part of a full intelligent manipulation system based on 3D vision

This paper shows how to extract information about the parts and their layout in a complex scene when a single range image is available. In the worst case, the complexity of the scene includes: no shape-restrictions, shades, occlusion, cluttering, contact, surfaces viewed in oblique angles and without textures. This work is a prerequisite before carrying out further robot interaction actions in the scene. The process is based on a novel 3D range data segmentation technique that avoids most restrictions imposed on other techniques. Making use of the 3D segmented parts, the method carries out an objects silhouette classification which allows us to perform a layout graph of the objects in the scene. A brief description of this method and experimental results are presented throughout the paper.

This paper is devoted to presenting a new strategy for 3D objects recognition using a flexible similarity measure based on the recent Modeling Wave (MW) topology in spherical models. MW topology allows us to establish an n-connectivity relationship in 3D objects modeling meshes. Using the complete object model, a study on considering different partial information of the model has been carried out to recognize an object. For this, we have introduced a new feature called Cone-Curvature (CC), which originates from the MW concept. CC gives an extended geometrical surroundings knowledge for every node of the mesh model and allows us to define a robust and adaptable similarity measure between objects for a specific model database. The defined similarity metric has been successfully tested in our lab using range data of a wide variety of 3D shapes. Finally, we show the applicability of our method presenting experimentation for recognition on noise and occlusion conditions in complex scenes.

A new 3-D data recovering system based on structured light-colour projection for active scenes is presented in this paper. With the use of a light projector and a single camera the system is capable of easily solving the correspondence problem when a 6-connectivity colour disordered pattern illuminates the scene. Since the pattern is handled as a set, the correspondence solution is based on set algebra operators. False matchings and errors are also detected and effectively solved. Experimentation and results show the goodness of the method.

We study a 3D visual servoing system composed of a robot manipulator, a CCD camera mounted on the end effector of the robot and a specific hardware. The use of computer vision as feedback transducer strongly affects the closed loop dynamics of the overall system, so that a visual controller is required for achieving fast response and high control accuracy. Due to the long time delay to generate the control signal, it is necessary to carefully select the visual controller. The objective of this work is to present a well structured design of efficient controllers to keep a desired relative 3D position between the object and the camera. Also an experimental setup has been built and used to evaluate the performance of the position based dynamic look and move system. Several results relative to different types of controller are presented. These results demonstrate the possibility of getting competitive results in real time performance with respect to more closed solutions shown in other works, while maintaining the advantage of easy system adaptability.

Presents a modeling strategy (half modelling wave, HMW) for the situation when only partial information of the 3D object surface is available. The method is applicable to unstructured range data. It means that no previous knowledge of the sensor used or the rough position of the object with respect to the sensor is required to compute the model. The developed partial modeling method has been conceived as a tool to be applied in new solutions for conventional 3D computer vision problems. Main contributions concerning partial viewing and 3D occlusion are also presented.

Presents a solution to the problem of determining the orientation of arbitrarily located 3D parts. The proposed algorithm is based on a new data structure specifically designed to store spherical representation models of 3D objects. This data structure called modeling wave set (MWS), allows keeping redundant information about local and global features of an object under a unique framework. It involves a different approach to deal with the modeling process providing new methods for solving generic problems like object recognition or pose determination. The algorithm described in this work is based on the fact that a rotation of a solid can be seen as a change in the wave used to model it, because of the invariance to spatial transformation of the MWS structure. This property makes an algorithm simpler and with lower computational cost than any other method based on spherical representations. The algorithm has been tested over a wide set of polyhedral and free-from shapes. Success ratio and 3D orienting errors obtained with the proposed algorithm can be considered as acceptable.