Logo multitel

Other technology and applications have been developed by the Image Processing department of Multitel. Indeed, thanks to our expertise in the field of image analysis and real-time embedded systems, we can offer research and development services in various areas of application.

For instance, the image processing department has conceived a prototype system to assist train backing for the SNCF (Société Nationale des Chemins de fer Français). This autonomous system is able to transmit high quality video images between the rear of a train and the operator in the driving cab. A wireless link robust over several hundred meters and able to function in very complex environments has been integrated in the prototype. Image and sound are in this way presented in real time to the operator with the help of a portable and autonomous device.

With its experience and expertise, Multitel is a highly skilled partner to meet research and development needs in the fields of analysis and transport of video and images.

Machine vision for:

  • Engineering integrated vision-based systems to improve quality, operating efficiency and safety (of both products and processes)

Application domains:

  • Automated digitizing and Archives and records digitizing and analysis
  • Automated inspection and control and manufactured objects (defects in products) & natural objects inspection (e.g. fruits grading)
  • Remote sensing and video streaming (surveying by satellite/UAV…)

Use cases:

  • Defects detection in windshield production chain (quality inspection)
  • Defects detection in glasses production chain (quality control)
  • Production chain monitoring (frictions and “hot points” monitoring)

Achievements

  • Automated inspection of railway traffic recordings

For the 9000 train accidents reported each year in the European Union, the Recordings Strip (RS) and the Filling-Card (FC) represent the only usable evidence for French authorities. Indeed, the RS contains information concerning the journey of the train, speed and related Driving Events (DE). The FC gives details on the origin/destination stations. A complete checking for 100% of the RS (instead of the 5% currently performed) was recently voted by French law enforcement authorities, which raised the question of an automated and efficient inspection of this huge amount of recordings. To do so, we propose a machine vision prototype, constituted with cassettes receiving RS and FC to be digitized. Then, the video analysis module determines the type of RS among eight possible types, time/speed curves are extracted to estimate the covered distance, speed, and stops while associated DE are finally detected using convolution process.

Automated inspection of railway traffic recordings use case – SNCF-BG

  • System to assist train backing

Multitel has been chosen to develop, test and validate a machine vision prototype to assist train backing, which is of primary importance for the French Railway Company (SNCF). In fact, the backing maneuver is a tricky operation for the train driver because he is not able to see the front of the freight train, which can easily reaches 700 meters long. In this case, the train driver is assisted by an operator in order to drive the locomotive which will push the train inside the hangar. The communication between both operators is actually done by radios, the operator is located at the front of the train, whereas the train driver is in the rear cab of the train, ready to drive. Currently, the operator takes important risks for visualizing the front of the train and maintaining a stable radio communication.

System to assist train backing – SNCF-ARV

  • Non-destructive quality control in laser welding process

This research project, called NDTLaser, aims at studying and developing a non-destructive quality control system, integrable inline inside the micro-welding process. The aim of the system is to monitor the welding process, based on the acquisition of signals coming from different sensors, such as cameras and microphones. Then these signals will be analysed using signal/image processing algorithms in order to detect potential defects and thus to determine the quality of the weld.

Print
rss2_feed