Industrial Collaborations and transfers

CEA List: Clickrog: Object grasping for disable persons

This contract started in November 2005. It is also supported by the Brittany Council through a grant to Claire Dune for her PhD (``krog'' means grasping in the Breton language). It is dedicated to object manipulation by visual servoing. The goal of this project is to allow disable persons to grasp an object with the help of a robotic arm mounted on a wheel chair. This task should be achieved with a minimum of a priori information regarding the environment, the considered object, etc.

France Telecom R&D: Cifre convention

This contract is devoted to support the Cifre convention between France Telecom R&D and Inria regarding Fabien Servant's Ph.D. The goal of the Ph.D. is to enable augmented reality on mobile devices like GSM or PDA used by pedestrians in urban environments. More precisely, its aim is to compute the absolute pose of the camera to show to the end-user geolocalized information in an explicit way.

France Telecom R&D: Augmented reality in urban environment

This contract started in October 2005. Considering the algorithm Markeless, the goal of this project is to evaluate its capability to handle rough models provided by geographic information systems. A second part of this project will consider the automatic initialization of the tracking process.

Past industrial Collaborations and transfer

Total Immersion

Software transfert: Marker-based augmented reality kernel

The Marker software implements an algorithm allowing the computation of camera attitude the calibration of the cameras using fiducial markers. Pose computation is handling using the virtual visual servoing approach. The idea consists in regarding the pose and the calibration as the dual problem of visual control. This method presents many avantages: precision identical to the traditional non-linear minimization methods, simplicity, effectiveness.
Licence: A licence of this software was yielded to the company Total Immersion.
Demonstration:: Virtual Visual Servoing: A framework for augmented reality

RIAM Sora Project

The goal of the SORA project (Objects tracking for augmented reality) is the developpement of tracking algorithm for augmented reality application. Our partners are Total-Immersion and VideoMage. Augmented Reality has now progressed to the point where real-time applications are being considered and needed. At the same time it is important that synthetic elements are rendered and aligned in the scene in an accurate and visually acceptable way. In order to address these issues a real-time, robust and efficient 3D model-based tracking algorithm is proposed for a 'video see through' monocular vision system. Virtual objects can then be projected into the scene using the pose. The tracking rate is 50Hz.

Cemagref-Ofival grant

Quality evaluation of porc meat

This study relates to the evaluation of the quality of sides of meat of pig by active vision and IRM measurements. It follows upon a joint proposal of CEMAGREF Rennes, Olympig company and Vista project at Ofival which agreed to give a financial support. Within this project, our task deals with a vision-absed estimation of the volume of piece of ham. This is join project with Cemagref for the ofival. We proposed algorithms for 3D reconstruction and exploration of a scene using a mobile and controlled camera. We used a space carving algorithm to obtain a precise and robust reconstruction of the 3D structure of unknown objects. To ensure the complete reconstruction of all the objects of the scene we present a gaze planning strategy that mainly uses a representation of known and unknown areas as a basis for selecting new viewpoints. The trajectory that leads to this exploration is handled using a visual servoing approach.

Ifremer

Dynamic stabilization of a pan and tilt camera for sub-marine image visualization

One major problem of under-water observation with an automatic engine is the instability of image acquiring. Indeed, this kind of small engines are submitted to low-frequency motions due to weak friction and currents of water. In this study, we proposed to maintain stabilization in the image by controlling the pan and tilt motions of the camera mounted inside the engine, using techniques applied for target tracking. The main idea within this approach lies in the fact that, as it is very difficult to track a point of an unknown scene using geometrical tools, it can be retrieved by the integration of its speed. Indeed, the velocity in the image can now be estimated in real-time, and without any a priori knowledge of the image content. Our approach has been validated on a dry set-up.

Controlling an uninstrumented manipulator by visual servoing

In this study, we proposed a method to control the displacement of a robot arm (the Victor 6000) with no proprioceptive sensor. The joint positions are not available and this manipulator is usually open-loop controlled. In order to get a more efficient control interface, we propose a closed-loop system based on an eye-to-hand visual servoing approach. We show that, using such an approach, measurement of the manipulator motion with proprioceptive sensors is not required to precisely control the end-effector motion. We propose solutions for position-based control and velocity control of the manipulator. To maintain the end effector in the camera field of view, the camera orientation is also controlled.