Saturday 5 January 2013

Automatic Empty Container Verification System

                   Introduction


Many inter-modal terminals expend significant resources verifying that shipping containers designated “empty” are actually empty at terminal gates. Typically, truck drivers park in the gate lanes or in a dedicated inspection area, exit the truck, and open the rear doors so that an inspector near the truck or at a central video station can visually verify that the container is empty. This process can take a minute or more per truck,wasting hours of gate and inspector time every day and putting drivers and inspectors at risk in traffic.
We describe and show performance results for an automatic empty container verification system (patent pending) which provides a faster, safer solution. The system automatically scans closed containers as trucks drive through, detects objects inside, and alerts inspectors for non-empty containers. Gamma-ray scanning technology images the interior of closed containers moving at up to 10 mph (16 km/h) in 5–10 seconds. Automated software analyzes the container image to detect objects inside the container, including pallets and other wooden objects, cardboard boxes, even plastic wrap. When an object is detected, it issues an alert to inspectors or the terminal’s information system. The system has demonstrated a probability of detection of about 97%.


Configuration
The system uses a Cs-137 source of photons. The gamma-ray energy of 662 keV provides good contrast for a wide range of materials, including small amounts of wood, cardboard, etc. The detector array is a single column of relatively large scintillation photon detectors to enable efficient photon detection and better counting statistics.
A change from the COTS version of the Portal VACIS is in the beam angle. COTS Portal VACIS systems are intentionally designed to offset the collimated source beam angle at about 10 degrees from perpendicularity to allow an oblique view through front and back walls of a container. This was a feature requested by operators since this is a favored location for false compartments, etc. and the oblique angle-of-incidence makes such compartments easier to detect. However, it was easier to develop automated algorithms to reliably detect front and rear walls with a truly perpendicular beam. This also significantly improves the ability to detect small objects placed against these walls.
A single operator’s booth can be used to monitor several lanes of outbound traffic. This is because the image analysis software operates essentially autonomously, and only requires operator intervention in the rare cases when the algorithm is uncertain of its results or if a container is actually non-empty.

No comments:

Post a Comment