I am seeing more and more quality control and machine vision applications in Intelligent Transportation Systems (ITS), where automated inspection of vehicles and their contents is performed for border control, port security and rail inspection applications. Much of this feels like factory automation, although instead of widgets and conveyors we have vehicles and lanes. There are some key areas of overlap between machine vision and ITS that have both similarities and differences. These include:
Acquisition: Lighting is a problem in outdoor applications. This often suggests use of cameras with auto-iris lenses or even day/night modes.
Specialty Illumination: Specialty lighting products are common in machine vision, and a huge need for specialty products designed for ITS is emerging.
Triggering: When does the vehicle show up? Real-time video-based triggering is an attractive alternative here, but requires hard-core machine vision software. Over the broad range of lighting and weather situations, this can be more challenging than it appears.
Velocity Detection: In law enforcement, traffic flow monitoring and vehicle characterization, velocity measurement is a critical parameter. Another hard-core machine vision tool, real-time pattern recognition, can be used to measure object motion in successive frames. This can be used to compute velocity.
License Plate Reading (LPR): LPR is a standard ITS component. Machine vision optical character recognition (OCR) algorithms are a good start, but LPR is a much harder problem. There is often little control of illumination, and the location, scale and angle of presentation of the plates is often highly variable.
Check out my article on the subject for more detail, and feel free to e-mail me at firstname.lastname@example.org with inquiries or ideas.