AVEN ENERABLE TECHNOLOGY
On its own or integrated with other technologies, OCR can enhance efficiency and security in intermodal operations. Terry Gibson of SAIC explains.
OCR systems can read numbers on a variety of equipment, including containers, chassis, number plates and gensets. Although OCR isn't a perfect means of identifying equipment, it's the best option available today. Printed identification numbers are currently the only standard identifier for many types of equipment. Until containers, lorries and other equipment are widely equipped with radio frequency identification (RFID) or some other reliable electronic identification method, printed numbers will probably remain the standard-and OCR will remain the most efficient method of recording them. A good OCR system can also improve data accuracy, recording equipment more accurately than human clerks in some situations.
An OCR system is essentially a data-collection system, the first link in the data-processing chain. Because of this, it can often be implemented fairly easily, with little modification to the rest of the chain. At the same time, since OCR data can support new productivity and security applications, implementing OCR can prepare a facility for further improvements down the road.
Conceptually, an OCR system is simple: video cameras at a checkpoint take images of equipment as it passes; computer software analyses the images to read the numbers; and the numbers are transmitted to the facility's information system. OCR systems are most often found at facility gates, but they can be installed at other locations to identify equipment almost anywhere in the facilityeven on quay cranes to identify containers as they are moved.
Of course, nothing is simple in the real world. Containers and lorries lead a hard life, and ID numbers can be damaged or obscured by weather and dirt, scrapes and scratches. Other unfavourable conditions like difficult lighting or poor contrast between numbers and background can compound the problem. OCR systems typically take multiple images of each piece of equipment-including both sides and the back of each container-to be sure to get at least one good clean image.
Because of difficulties like these, no OCR system will read every number correctly. The facility needs a way to handle exceptionstypically a clerk who views an image of the equipment on a video display and updates the number in the system if required. To help the clerk, the OCR system may raise a flag for "low confidence" numbers, cases where the system had trouble determining the number.
OCR accuracy is a measure commonly used to evaluate OCR systems. It's a useful measure, but potential buyers need to be cagey about accuracy claims. A supplier who claims to read the numbers on 95% of all containers is going to have trouble living up to it, if only because many container numbers are in such poor shape that even a human can't read them. A more realistic measure of accuracy is the percentage of machine-readable numbers-excluding numbers that are badly damaged, obscured, in non-standard format, and so on.
Typically, about 10% of containers are not machine-readable, but this figure varies depending on the condition of the containers in a facility's actual traffic. When a supplier says his OCR system reads 95% of machine-readable containers, he's actually saying it will read about 85 percent of all container traffic if 10% are not machine-readable, but about 90% of all container traffic if 5% are not machine-readable.
Although the supplier may seem to be inflating his accuracy by tossing out non-machine-readable numbers, it gives the buyer a better idea of the accuracy he will get with actual traffic.
Depending on many factors, an OCR system can pay for itself in two or three years just by streamlining the equipment identification process. And buyers can recover the cost even faster by using the system to support other applications. A good example is online damage inspection. At many facilities, inspecting containers for damage is done manually, with inspectors walking around each container at a checkpoint and noting damage on a clipboard or handheld data-entry device-a time-consuming approach. The images from an OCR system can be used to streamline the damage-inspection process. Special line-scan cameras can provide extremely high-quality colour images of containers for both OCR and damage inspection. A clerk at a damage inspection workstation in a central location-on site or thousands of miles away-can view the images of any container, either live, as the container is handled, or from an image archive, long after the fact, and note damage directly into the system.
And damage inspection is just the beginning. OCR can be an important component of larger-scale integrated systems that enhance both efficiency and security. For example, SAIC's Intelligent Intermodal Solutions (IIS) combine OCR and other technologies to support processes at intermodal facilities. The following illustrates how an IIS system might support an efficient, secure process for identifying inbound lorries.
1. At the in-gate, a lorry drives through an IIS portal, where highresolution cameras take images of the number plate and container.
IIS uses OCR to read identify the lorry and the container. The images (for damage inspection) and numbers are added to a transaction record in the IIS database.
2. The lorry stops at an IIS kiosk in an inbound lane. The kiosk uses a smart card, biometrics or other technique to identify the driver. At the same time, IIS uses OCR to read the number plate again. IIS uses the plate number to find the transaction record created at the portal and adds the driver data to the record.
3. IIS sends the equipment and driver data to the terminal operating system (TOS). Using this data, the TOS authorises the lorry to enter the terminal and issues instructions to the driver through the kiosk. The lorry proceeds into the terminal.
At the out-gate, IIS performs a similar process to confirm that the lorry is carrying the correct container out of the terminal. Both processes are performed automatically, with only one brief stop at the kiosk.
In another example, SAIC's developmental Integrated Container Inspection System (ICIS) combines OCR with high-throughput gamma ray imaging and radiation scanning to help terminals provide security data for Customs agencies and other security authorities. ICIS scans and identifies high volumes of containers in the flow of terminal traffic and integrates the images and data from these and other sources (such as manifests and watch lists) in a central database. Authorities can access and review the images and data for any container in seconds, thus enabling them to perform a first-stage examination without holding the container. In this way, the integrated system benefits the authorities by providing useful, timely data, and benefits the terminal by supporting the collection of security data with minimal impact on operations.
In summary, OCR is a venerable technology with a bright future. On its own, OCR can streamline the processing of containers and other equipment with relatively little overall impact on operations. And OCR can be combined with other technologies in sophisticated systems to enhance both productivity and security in intermodal operations.
Terry Gibson is vice president of corporate development in SAIC's Security and Transportation Technology Business Unit where he guides the development of cost-effective, integrated systems to enhance security and productivity throughout the global intermodal industry."
LATEST PRESS RELEASES
The exhibition series ‘Intermodal Africa’ organized by Transport Events is always a good possibility... Read more
The German based Headquarters of ShibataFenderTeam recently completed an order for the Port of Esbje... Read more
Start-up for ocean route planning enters ESA’s Business Incubation Centre (BIC) Read more
AMRO, a specialist marine equipment and services provider, is proud to announce that they will now c... Read more
Ninth Consecutive “Excellent” Coast Guard Security Assessment Awarded to Port of Baltimore Read more
In 2017, the US office of ShibataFenderTeam delivered 55 nos. CSS 1450 Cell Fender Systems (G2.0), 8... Read more