The unique screening solutions helping solve the unique challenges in the fight against illegal drug trafficking
By Thierry Ball, Business Development Manager Ports & Borders, Smiths Detection
The uniqueness of the challenge
The global effort to combat drug trafficking across land, sea and air borders involves numerous entities working towards one common goal – customs, border control agencies, and more recently also courier services as traffickers exploit the surge in e-commerce and use freight forwarding and mail delivery companies to transport narcotics.
The detection of these narcotics using X-ray technology faces distinctive challenges, however. To either stretch the product to increase profits or generate a different effect on the user, they are often mixed with other substances, which changes their physical characteristics, and are inconsistent in shape and form. This makes narcotics detection more complex, especially compared to the very regulated detection of explosives, which must have certain known characteristics to “work” and be an actual threat.
Thanks to major technological developments, there are now several highly effective solutions available to help tackle this challenge:
- Augmented X-Ray (AXR) algorithms for High-Energy X-ray Transmission based non-intrusive vehicle inspection systems.
- Advanced object recognition and material discrimination algorithms augmenting conventional X-ray and Computed tomography (CT) scanners for air cargo and baggage screening.
- X-ray diffraction (XRD), a well proven and established method for material-specific detection based on molecular structures.
Augmented X-ray: A quantum leap in screening cargo and vehicles
Non-intrusive inspection (NII) systems, using High-Energy Transmission based X-Ray, are traditionally being used for cargo and vehicle screening to intercept contraband at land and sea borders. These systems can now be taken into a completely new era by deploying Augmented X-ray (AXR) algorithms.
Augmented X-ray, for example, has the capability to identify overlapping objects of different materials (and therefore areas with potential anomalies) on the X-ray image. By ‘de-overlapping’ or removing material, it can re-calculate the hidden substance, discriminate more accurately, and display more relevant information on the material.
The smart algorithms can also remove steel container or vehicle walls from the X-ray image to highlight organic materials throughout the entire screening area irrespective of depth location and thereby significantly support operators in detecting concealed narcotics.
These advances are made possible by capturing substantially more data using small-pitch detectors and enable new image processing algorithms. Critical information can be extracted from the data to support fast and accurate image analysis. Operators can focus on specific areas of interest, spotting anomalies and hidden items quickly and efficiently.
Harnessing the power of CT for parcel and baggage inspection…
CT scanning is a sophisticated imaging technique that employs X-rays to create detailed three-dimensional images of objects, which enable operators to make fast and accurate judgements on the content of a bag or a parcel.
It has also proven to be very effective in automatically detecting explosives and more recently also contraband, prohibited items, dangerous goods, and other targeted objects in parcels and bags.
When it comes to automatic narcotics detection, CT scanners have two powerful tools in their arsenal:
- material discrimination and
- object recognition algorithms.
Both come with their individual benefits and challenges in terms of narcotics detection.
Material discrimination relies on image processing to segment and classify items based on X-ray absorption characteristics, such as density or effective atomic number (Zeff). On this basis, it then leverages machine learning to determine if a threat is present or not. This approach is exemplified in automatic explosives detection.
While the further development of algorithms continuously improves the detection capabilities, it is most effective for substances with known physical characteristics and can therefore detect relatively pure narcotics with a high degree of probability.
Object recognition, on the other hand, leverages deep learning to develop algorithms that mimic the human brain’s data processing. These algorithms learn to identify patterns in the shape and texture of objects by analysing thousands of X-ray images. While material characteristics are considered, the focus is primarily on colour and intensity in the X-ray image, such as identifying metal in blue and organic material in orange.
Several smart, adaptable deep learning algorithms are in common use for detecting dangerous (e.g. lithium batteries) and prohibited goods, achieving high detection rates and low false alarm rates.
They can also be used to identify pills and blister packs based on their shape. The optical signature is however often similar to benign tablets. In addition, substances lacking a distinct shape, like powders or liquids, pose difficulties for this technique.
In essence, CT technology relies on either clear physical characteristics or a distinct shape to automatically detect narcotics. X-ray diffraction adopts an entirely different approach…
…And enhancing it with the power of XRD
XRD is a technique that exploits the interference patterns generated when X-rays interact with the molecular structure of materials. It is well-established in many industries and recent technological advances now make it a viable option for customs screening as well.
Each atom inside a molecule scatters the incoming X-ray waves. Depending on the spacing between atoms, different interference patterns occur. The resulting signals offer insights into this spacing. As each substance possesses a unique spacing pattern, XRD employs these measurements to generate a ‘diffraction fingerprint,’ facilitating differentiation between materials, even those with closely matched densities.
By comparing these patterns with a database of known narcotics, authorities can determine the composition of suspicious substances. This precise approach makes it particularly suitable for detecting constantly changing powder, liquid or solid compounds such as narcotics.
Collaboration is key
Collaboration with authorities, along with the extensive collection of samples and images of substances to train algorithms, supports the continuous development of increasingly effective solutions.
To this end Smiths Detection, a developer of smart learning algorithms in security screening, is part of an EU consortium to develop an intelligent customs border control screening system (BAG-INTEL) that will automatically detect narcotics in passenger baggage. The BAG-INTEL consortium brings together technology providers and customs agencies to solve customs and border control challenges by targeting specific bags needing manual inspection. The algorithm will be developed by leveraging a combination of material discrimination as well as shape-based, deep learning.
Where we go next
Recognising that no singular security screening method can guarantee 100% interception of all threats and contraband across the variety of border crossing points and transportation routes is crucial. Deploying adapted means of screening using multiple complementary scanning techniques is the way forward to combat drug trafficking.
The answer for detection, therefore, involves employing a blend of NII vehicle inspection systems, CT screening and XRD scanners, enhanced with smart detection algorithms.
Ongoing research and development efforts aim to improve sensitivity, reduce scanning times, and enhance the overall capabilities of these technologies. Integration with artificial intelligence and machine learning also play a crucial role in refining narcotics detection processes.
Although object recognition and material discrimination can be used successfully together, it is likely that, in the longer term, XRD will provide highly accurate information about the physical composition and would be best suited for automatic narcotics detection.
In the relentless pursuit of curbing narcotics trafficking, CT and XRD have established themselves as indispensable tools. Their unique strengths, when integrated, offer a solution to the challenges faced in narcotics detection. As technology continues to advance, the collaboration between CT and XRD promises a more accurate, efficient, and sophisticated approach, ultimately contributing to a safer and more secure global community.
Ultimately, we are likely to see CT and XRD technology combined into one unit. In this scenario, each item will be checked by both technologies simultaneously, leveraging the strengths of both techniques. This will add another line of defence for Customs agencies and security operators alike when coupled with existing technologies.
