�Ӱ�ԭ��

For Michelle Fabiani, Ph.D., assistant professor of criminal justice at the University of New Haven and co-director of the , her work examining looting activity at archaeological sites has spanned more than 15 years. With the support of a Faculty Research Project grant from the CT Space Grant Consortium, she aims to better understand the environmental and human factors driving this illicit activity, providing critical insights to protect these historical sites.

�Ӱ�ԭ��Archaeological looting is a persistent phenomenon around the world,�Ӱ�ԭ�� she explains. �Ӱ�ԭ��And as archeologists typically only work a couple of months a year, we likely know only a fraction of what happens when they are not there.�Ӱ�ԭ��

Now, Dr. Fabiani is leading an initiative to monitor 661 archaeological sites in Lower Egypt over a period of eight years. The project uses satellite imagery and data-science techniques to detect changes on the ground, offering insights into how looting attempts fluctuate over time.

�Ӱ�ԭ��The goal is to develop indicators to determine which types of sites are most likely to experience looting activity in response to different factors �Ӱ�ԭ�� social, political, environmental, and economic,�Ӱ�ԭ�� Dr. Fabiani says.

�Ӱ�ԭ��We�Ӱ�ԭ��re using algorithms to detect potential looting pits from satellite images�Ӱ�ԭ��

The work builds on a previous pilot project in which Dr. Fabiani hand-coded over 1,000 images across a subset of 140 sites in Lower Egypt. This new project expands the scope and incorporates advanced algorithmic-detection methods, in collaboration with Dr. Shivanjali Khare and students from data science and criminal justice.

�Ӱ�ԭ��We�Ӱ�ԭ��re using algorithms to detect potential looting pits from satellite images,�Ӱ�ԭ�� she explains. �Ӱ�ԭ��This will create a robust dataset and method for identifying attempted looting activity on a much larger scale than before.�Ӱ�ԭ��

Dr. Fabiani�Ӱ�ԭ��s work is rooted in the recognition that attempted looting is connected to larger global forces, and she has . �Ӱ�ԭ��Objects of antiquity tend to have inherent monetary and aesthetic value,�Ӱ�ԭ�� she explains. �Ӱ�ԭ��They�Ӱ�ԭ��re excellent forms of collateral in illicit economies because they�Ӱ�ԭ��re untraceable.�Ӱ�ԭ��

The values of these artifacts on the art market rarely, if ever, experience crashes. This makes antiquities attractive to those looking for a stable, high-value commodity.

The project will help uncover patterns and identify risk factors, much like how meteorologists predict storms or earthquakes. �Ӱ�ԭ��We want to generate data that can be used to proactively manage heritage sites,�Ӱ�ԭ�� Dr. Fabiani says. �Ӱ�ԭ��It can give governments and local authorities a chance to more proactively allocate resources for protection.�Ӱ�ԭ��

�Ӱ�ԭ��We�Ӱ�ԭ��re trying to scale up what has been done before�Ӱ�ԭ��

Dr. Fabiani�Ӱ�ԭ��s project is ambitious, spanning eight years and hundreds of sites. The research is not only expanding what we know about attempted archaeological looting but also how data science can play a role in heritage management.

This interdisciplinary approach �Ӱ�ԭ��combining criminal justice, data science, and archaeology�Ӱ�ԭ�� has opened new doors for understanding looting activity.

"By incorporating change-detection algorithms, we can automate the process of identifying potential looting activity in satellite images."Dr. Fabiani

At a scale of more than 600 sites to monitor, relying on human coders manually scanning satellite images becomes impractical. �Ӱ�ԭ��We�Ӱ�ԭ��re trying to scale up what has been done before,�Ӱ�ԭ�� she explains. �Ӱ�ԭ��By incorporating change-detection algorithms, we can automate the process of identifying potential looting activity in satellite images.�Ӱ�ԭ��

By automating the process, much more ground can be covered in less time, increasing the efficiency and scope of the research. �Ӱ�ԭ��We work better when we have diverse skillsets that speak to each other,�Ӱ�ԭ�� she says.

�Ӱ�ԭ��It�Ӱ�ԭ��s about being able to say, �Ӱ�ԭ��Here�Ӱ�ԭ��s where you need to focus your efforts�Ӱ�ԭ���Ӱ�ԭ��

Satellite imagery is not always available, and when it is, the resolution may not be high enough to confirm that a small disturbance in the ground is indeed a looting pit.

�Ӱ�ԭ��Sometimes, we�Ӱ�ԭ��re looking at one or two pixels, and it could just as easily be a bush or a shadow,�Ӱ�ԭ�� Dr. Fabiani admits. �Ӱ�ԭ��It can be difficult to confirm what we�Ӱ�ԭ��re looking at, especially with historical imagery.�Ӱ�ԭ��

�Ӱ�ԭ��We must come up with other ways of validating what we�Ӱ�ԭ��re doing,�Ӱ�ԭ�� Dr. Fabiani continues, �Ӱ�ԭ��which comes down to more data and computer-science approaches, as well as cross-validation with multiple coders.�Ӱ�ԭ��

Looking ahead, Dr. Fabiani envisions a future where the data from this project can be used to inform proactive heritage-management strategies, allowing governments to allocate resources more effectively. �Ӱ�ԭ��It�Ӱ�ԭ��s about being able to say, �Ӱ�ԭ��Here�Ӱ�ԭ��s where you need to focus your efforts to achieve the greatest results,�Ӱ�ԭ���Ӱ�ԭ�� she explains.

Her work offers a glimpse into the future of heritage protection. With the right infrastructure and continued support from grants such as the one from the CT Space Grant Consortium, the project could help create a more informed and strategic approach to preserving shared history. �Ӱ�ԭ��This method could be replicated in many regions�Ӱ�ԭ�� Dr. Fabiani suggests, �Ӱ�ԭ��and used to develop policies that continue to help protect cultural heritage on a global scale.�Ӱ�ԭ��