X-ray vision

When passing through a baggage screening zone, visitors to buildings and travellers struggling with their luggage at airports will probably give little thought to the challenges involved for the X-ray operators. They probably think they could just sit down and perform the task very easily themselves with little or no training to speak of. That would be a serious underestimation of the demands inherent in the role.

Primarily, X-ray system operators need excellent powers of concentration and the ability to focus their attention consistently and appropriately. They must also absorb and make use of a considerable amount of information concerning threats, mentally translate what are not particularly realistic representations of bags’ contents and subsequently take totally reliable decisions about whether suspicious items are present.

The challenges are increased with the expectations of visitors and travellers, as well as commercial imperatives (meaning that bags have to be screened with minimal delay).

Pressures are also increasing in certain sectors. The UK’s airports already handle over 180 million passengers every year. The Department for Transport is currently predicting that the number of UK air passengers will increase by 25% during the period 2000-2005, and by a further 25% between 2005 and 2010.

Also, market conditions determine that pay rates for X-ray operators aren’t that high. Considerable staff turnover is the end result, which may well impact on the overall calibre of employees and their experience levels.

Making screening effective

Viewed in this context, what issues need to be considered in maximising the effectiveness of the screening process? Coming up with the right solutions is dependent on having a real understanding of the problems that operators actually face when looking at an X-ray image. If a knife is wedged between two items of clothing, with its edges facing the bottom and top of the case, what will it look like to the security operative? It will appear as a thin sliver which is exceedingly difficult to interpret.

Similarly, if a bullet is close to a perfume bottle or a hairdryer, it may well appear to be a part of it. In 2D it would be difficult to assess whether the bullet was in fact above or below the other object.

Between 1993 and 1995 Dr Gerald Gibb of Galaxy Scientific carried out some research to establish how operators interpreted images, and what caused them difficulties in doing so. Gibb established that the main factors are:

• clutter – this makes it harder to assess what’s actually in a bag or suitcase;
• resolution, as the sharpness of the pictures is hugely important in assisting interpretation;
• opaque objects (because they block the immediate view);
• electronic items – they’re quite complex and can add clutter;
• the development of improvised explosive devices (IEDs) – increasingly, these are split into component parts and don’t have a typical ‘bomb-like’ appearance;
• problems in interpreting 2D images (they don’t render a realistic representation of objects in their surroundings, and offer no depth clues at all).

Testing times for operators

Given the very specific skills required for the job, it’s useful for security managers to have in place some kind of assessment of individual aptitude for system operatives that goes way beyond the obligatory application form and interview. For example, tests can seek to establish candidates’ ability to distinguish figure from background or to detect threats prior to any training.

In the UK, the Department for Transport recommends psychometric tests to be used by security suppliers. Practical difficulties in carrying out testing for all candidates right across the security sector relate to the fact that the pool of potential employees may be small, making it problematic – in some cases – to justify going beyond testing for characteristics that would make the role impossible to fulfil (such as colour blindness).

Clearly, quality training is absolutely crucial in elevating operators to the right standards of performance. Training is required by legislation for both the aviation and the prison sectors in the UK, and their X-ray system operators must also pass the post-training test prior to taking on a screening role.

Areas covered by training should include how to operate the machines, keyboard functions (and the significance of colours projected by the equipment), how typical prohibited items will appear and what procedures to follow if a threat is detected.

It’s worth bearing in mind here that a considerable focus on IEDs and component parts of electrical items is vital because both are difficult to interpret, and the former may be hard to spot. Don’t forget that refresher training is also needed to keep the skills and knowledge of operators fine-tuned and up-to-date (aviation screeners working in the UK are required to complete repeat training once every year in any case).

Another key issue in performance terms is the ability of the operator to concentrate. Considerable research on this topic has shown that vigilance drops off after the first few minutes, not to mention the health dangers resulting from staring at a screen for long periods. UK-based aviation authorities require screeners to switch over to other duties after only 20 minutes. It’s counter-productive – and against Health and Safety requirements – to expect security personnel to undertake screening duties for extended periods of time whatever the environment and circumstances.

Distraction is obviously an important problem. X-ray system operators can momentarily lose focus if their eyes are drawn to something away from the screen, or if there’s too much ambient noise. This should be highlighted during training, and operators encouraged to develop strategies for dealing with these situations (for example, by stopping the conveyor belt if something or someone is creating an unwanted distraction).

A major software development – designated ‘threat image protection’ (TIP) – has a highly influential role to play in overall operator performance. Currently used in the main by the aviation and prison sectors, TIP offers a library of thousands of images which represent the way in which potential threats have previously been presented.

During the course of the working day, TIP software periodically and randomly simulates the type of threat that the X-ray operator would be expected to spot, with their 'score' providing a true and objective measurement of overall performance

During the course of the working day, TIP software periodically and randomly simulates the type of threat that the X-ray operator would be expected to spot, with their ‘score’ then providing a true and objective measurement of overall performance. This is extremely useful as an ongoing training process. It also enables analysis of performance in relation to apparent aptitude before taking on the job – very useful in the refinement of selection procedures – and analysis of that performance at different times of the day.

In addition, it helps to maintain the operators’ interest and motivation. Otherwise, they may become bored due to the routine of the job and the relative infrequency of actual threats being presented. Sometimes, suppliers place considerable emphasis on the wide variety of keyboard functions available. In my experience this isn’t necessarily a benefit, as operators have often failed to come to terms with all of the keys on a fairly standard keyboard and tend to stick to those with which they’re most comfortable.

There are eight function keys on a typical X-ray machine. These enable the operator to differentiate between organic and inorganic material (as the materials are shown in different colours), enhance low or high density objects, reverse colours on the screen (this can sometimes help in being able to see particular types of object more clearly) and change from colour to black and white (which can also help in identifying particular items).

In addition, operators are also able to zoom in on particular areas of the screen so that they can be viewed in more detail, and ‘flip’ images to see items that would otherwise be hidden.

Image diagnoses: the key points

To achieve optimum results, operators need to have a systematic way of diagnosing the image, which includes knowing the appropriate circumstances in which to use each key.

For example, if they can see two wires but aren’t sure where they’re going, they can change the image to black and white and then look at the negative. The wires should appear bright white, and it will be easier to see where they lead. If an operator wants to look at an electronic device more closely, they could use the inorganic stripping function to enable them to focus on any organic material that remains. If orange matter is revealed inside that isn’t perfectly rectangular then this would merit further investigation... it could well be explosive material.

Interpreting 2D images places quite a cognitive load on the screener, whereas new 3D technology provides a much more realistic representation of objects. 3D vision is created in normal circumstances by humans using both eyes to capture two slightly different views of an object. The human brain then adopts a process of triangulation – otherwise referred to as ‘binocular parallax’ – to calculate the position of the object in 3D space.

The new technology mimics this in a process called ‘stereoscopic viewing’, which involves capturing two images of an object from two slightly different points of view. Psychological depth cues in the human brain are then activated by rapidly alternating both images on a special display to give the viewer the perception that objects extend both in front of and beyond the screen. Less time is spent making sense of the overall image, such that more time can be devoted straightaway to analysing individual items and patterns.

By making the scanning process more straightforward for the operator, this technology assists in threat detection and speed of throughput, while also reducing training time and the likelihood of false alarms.

Another important development of late has been the ability to enhance detail without zooming. In photographic terms, consider a photograph of Warwick Castle, where the building is still at some distance but you can see the brick detail. This is useful because zooming does reduce picture quality, and there can also be a tendency not to consider areas other than the zoomed section.

Will humans be made redundant?

Clearly, X-ray system operators do face significant challenges. There’s considerable scope for human fallibility. That being the case, would security managers benefit from the process becoming fully-automated and removing trained security operatives from the equation altogether?

You really can’t see this being a viable option for some years to come. There are simply too many permutations in terms of what might turn up in peoples’ baggage. Human discernment is still needed to make sense of it all. Therefore, we’ll simply have to maximise the relationship between human and machine.

On the machine side, that will mean working to make life easier for the operator, and ensuring that technological advances really are practical for the human element.

In terms of the operator, the managerial challenge is to move them closer towards the virtuoso end of the scale in terms of their ability to work and interpret the system – which, after all, is totally dependent on aptitude, training and levels of job satisfaction.

Nick Fox is chief executive of 3D X-Ray Ltd (www.3dx-ray.com)