Automatic Recognition and Visual Inspection of Transportation Containers 

Automated transportation container handling makes container terminals and border control stations flexible and increase their speed and throughput. This is important in the world of ever-growing transportation traffic.

Visy Oy is a company specialized in machine vision technology. One of company’s main areas of interest is optical character recognition (OCR). Visy has developed a system that can be used for automatic recognition of container identification numbers and for visual inspection of containers and trailers. 

Container Imaging

Container lorries are imaged by using special cameras. These cameras enable imaging of variable sized containers with extremely high resolution and good image quality. The typical resolution of a lorry image of is around 2000x10000 pixels (height x width). This means that details even in millimeter class are observed. It is possible to image the sides as well as the roof of a container by using appropriate camera installation.

Lorries are imaged while moving. This means that the traffic do not have to be stalled due to imaging and therefore traffic flow will be smooth. The distortions caused by speed variations are removed from the image. As a result, a high quality image with natural proportions is obtained. The containers and the identification numbers are located and extracted from this image.

Visy container and license plate recognition system at Vaalimaa Customs station.

Container Recognition

Container recognition is based on digital image processing and neural network technology. The algorithms have been developed with robustness and universal applicability in mind. Adaptive algorithms are often used. Neural network technology enables teaching of new character sets to the system. All this adds up to a system that is versatile, globally usable and reliable, even in difficult weather conditions.

Before recognition, the images are enhanced. After that the separate characters and character groups are located. The found character images are then fed into neural network that classifies them into known character classes. The system recognizes both container chassis numbers and ISO codes.

Visual Inspection of Containers and Trailers 

Images and recognition event data can be stored to an internal database or given as output to an external data management system. The system provides tools for database quering and browsing. Combined with the high image quality, the visual inspection of containers and trailers is effortless. The tool for visual inspection with zooming properties and image enhancement capabilities is included in the system. The stored images also act as a good reference in reclamation situations.

Other Features

License plate recognition ability can be integrated as a part of the container recognition system to obtain information from the transportation vechile. Based on the container and license plate recognition results, the system can trigger alarms and send them to a browser or to an external system like TETRA mobile radio network.

Measurements like weight and dimensions of a container can be added to the recognition event data. Weight is received from an external weighing equipment. The dimensions can be measured from the image or by using external equipment.

Moreover, the system has remote use capability via Internet connection or phone line.

Experiences

The system has been tested succesfully in harbours in the Netherlands and in Finland. A complete container and license plate recognition system was installed at Vaalimaa customs station in the end of the year 2000. Also, Visy Container Imaging and Recognition system was awarded the third INNOSUOMI - prize in the year 2000.