Rail weighing systems
Viinex Foundation 1.4 has a comprehensive set of modules for integration and deployment with a weighing software. It is capable of railcar ID number recognition (for rolling stock in ex-USSR) for static and dynamic weighing, and solves most typical issues for this:
- Capturing video from several cameras around the railcar, optionally presenting the view to the operator and storing video to the archive;
- Performing railcar ID number recognition using video streams from both sides of a railcar to ensure the best results even if the ID number sign is contaminated;
- Performing number recognition using video stream from the side of a railcar as well as from its truck;
- Recognition results consolidation: the results of railcar ID number recognition from multiple cameras are automatically consolidated to obtain a single best hypothesis;
- Railcar separation and counting on video using an optical / IR sensor;
- The ability to control the railcar separation and number recognition from within weighing software to make sure that railcar list output by Viinex Foundation and by weighting software is agreed;
- Fisheye distortion compensation: in case if video camera cannot be installed within recommended distance from railway, lenses of short focal length have to be used in order to provide field of view which is wide enough. Such lenses usually introduce significant image distortion which can be compensated by a special module implemented in Viinex Foundation 1.4.
- Train manoeuvring support: if end user’s business process requires the train to move back and forth before stabilizing on a weigh bridge, Viinex Foundation 1.4 implements a special logic to assign correct order numbers to railcar ID number recognition results. This is achieved by installing three optical sensors together besides the video camera, to register railcar movements direction, and a software module which takes this information into account to output recognition results.
- Optical train movement direction detector: an optional module which implements a train motion direction estimation on a video stream, without additional sensors.
Vehicle weigh bridge
Automotive vehicle license plate recognition modules are also implemented in Viinex Foundation 1.4 for use together with software for auto weighs. The AutoLPR feature and accompanying modules in VF 1.4 have the following advantages:
- Viinex Foundation 1.4 implements all the work for interaction with video surveillance equipment. The developers working on weighing software or the integration of weighting system with an ERP don’t have to spend their time on video acquisition, storage, rendering and so on, - it’s all on Viinex Foundation.
- There are several LPR engines supported from multiple vendors, to choose the best one for the specific environment where the system is to be installed, for instance the prevailing type of license plates or prevailing country of origin (different vendors’ LPR engines perform differently depending on type of license plates to be recognized). What’s important, weighing software sees a unified programming interface provided by Viinex Foundation 1.4: there is no need to modify weighing software’s source code to switch between LPR engines from different vendors.
- License plate recognition is supported in both “checkpoint” and “free flow” modes, i.e. when a car stops and when it doesn’t within camera’s sight. This makes Viinex Foundation 1.4 suitable for both static and dynamic weighing.
- Viinex Foundation 1.4 supports simultaneous recognition of license plates at both front and back of the vehicle, which is implemented in two fashions: when these two plates are interpreted as one number, and then these are two different numbers. Results gathering is performed by so-called consolidation module. In the first case it uses results from the front and the back license plate to form the best recognition result. It can enhance the reliability of recognition using results from two license plates, just the same way as it’s done for railcar ID numbers. In the second case, when the license plates in front and in back of the vehicle are different, - which is typical for a truck and a trailer, - the results for one vehicle are automatically gathered together and reported to the weighing software in a single package. In both cases, the integration is made easier with consolidation module implemented in Viinex Foundation 1.4, because the weighing software needs only to know about a checkpoint where the recognition is performed, - not about the explicit video cameras.
- Flexibility of deployment: Viinex Foundation 1.4 has an interface library which connects to Viinex Foundation sever software. This makes it possible to deploy Viinex Foundation 1.4 on a separate computer, or place it in a virtualized environment, - or co-locate it with the weighing software,- depending on project requirements.