Posted in | News | Machine-Vision

Artificial Intelligence to Identify Bombs from Vietnam War

Artificial intelligence was recently used by scientists to identify bomb craters in Cambodia from satellite pictures. These bomb craters occurred during the time of the Vietnam War. The team believes that artificial intelligence can help detect unexploded bombs.

Bomb craters as seen by satellite. Image Credit: The Ohio State University.

The latest technique increased the detection of real bomb craters by over 160% when compared to that of traditional techniques. Integrated with declassified U.S. military records, the model proposed that 44% to 50% of the bombs located in the studied area might still remain unexploded.

According to Erin Lin, assistant professor of political science at The Ohio State University, current attempts to detect and securely remove unexploded bombs as well as landmines, known as demining, has not been as effective as required, in Cambodia.

In this context, Erin cited a new report that was commissioned by the UN. In the report, the Cambodian national clearance agency has been criticized for providing a picture of rapid development by targeting areas that have minimal or no threat of having unexploded mines. The report calls to focus on areas that are more highly dangerous.

There is a disconnect between services that are desperately needed and where they are applied, partly because we can’t accurately target where we need demining the most. That’s where our new method may help.

Erin Lin, Assistant Professor, Department of Political Science, The Ohio State University

The study was jointly headed by Lin and Rongjun Qin, an assistant professor of civil, environmental, and geodetic engineering at The Ohio State University. The study was published in the PLOS One journal.

For their research, the scientists began with a commercial satellite picture of a 100 km2 area that is close to the town of Kampong Trabaek located in Cambodia. This locality was the target of carpet bombing by the United States Air Force from the period of May 1970 to August 1973.

The scientists utilized a type of artificial intelligence known as machine learning, to study the satellite images for proof of bomb craters.

The study is significant because the scientists are now aware of the number of bombs that were dropped in the region and also the general spot of their fall. Through the craters, the scientists were able to know the number of bombs that had truly exploded and the location of the explosion. They can subsequently determine the number of unexploded bombs that are still left and the particular localities where they could be found.

Lin informed that the research involved a two-stage process. During the first stage of the process, the scientists utilized algorithms that were specifically made to identify meteor craters on the planets and moon. This allowed them to identify a number of potential craters, but still, it was not sufficiently good.

Lin added that bombs indeed create craters that are analogous to—albeit smaller than—those created by meteors.

But over the decades there’s going to be grass and shrubs growing over them, there’s going to be erosion, and all that is going to change the shape and appearance of the craters.

Erin Lin, Assistant Professor, Department of Political Science, The Ohio State University

The second stage of the procedure builds on the complexities of how meteor craters are different from the bomb craters. The computer algorithms considered the innovative features of bomb craters, such as their sizes, textures, colors, and shapes.

Once the machine “learned” the technique to identify true bomb craters, one of the scientists checked the work of the computer. Up to 177 true bomb craters were detected by the human coder.

While the first stage of the model developed by the scientists detected 89% of the real craters (157 of 177), it also detected 1,142 false positives—crater-like features that were not induced by bombs.

The second stage of the model removed as much as 96% of the false positives and lost only five of the true bomb craters. Hence, its rate of precision rate was approximately 86%, detecting 152 of 177 craters. Lin added that the proposed technique increased the detection of true bomb craters by over 160%.

Additionally, the team had access to the declassified military data that indicated that 3,205 general-purpose bombs, called carpet bombs, were dropped in the region examined for this research.

Integrated with the demining reports and the study results, this data, indicates that about 1,405 to 1,618 unexploded carpet bombs have still not been accounted in the region. That is roughly 44% to 50% of the bombs dropped in the region, stated Lin.

The study mostly covered agricultural land, which means local farmers face a greater risk from unexploded bombs, Lin stated. The threat is not hypothetical.

In the past 60 years, after the bombing of Cambodia, over 64,000 individuals have been injured or killed by unexploded bombs. Currently, the injury count averages to one individual each week.

The process of demining is expensive and time-intensive, but our model can help identify the most vulnerable areas that should be demined first.

Erin Lin, Assistant Professor, Department of Political Science, The Ohio State University

Other study co-authors were Jared Edgerton, a doctoral student in political science, and Deren Kong, a former engineering graduate student, both from The Ohio State University.


Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.