Crime scenes are usually available for investigation for only a short time.
During those crucial hours, police must carefully record every piece of evidence before the area is cleaned up or reopened. Even a small object or tiny detail can later become important in solving a case.
Now, researchers at the Technical University of Munich (TUM) and the Bavarian State Criminal Police Office (BLKA) are developing a new technology that could make crime scene investigations much smarter and more efficient.
The project aims to create intelligent 3D digital copies of crime scenes that investigators can do more than simply look at.
Instead of only viewing a virtual room, they will be able to ask questions and receive answers from artificial intelligence (AI), making it much easier to find important evidence.
Police in Bavaria already use detailed 3D models of crime scenes. These digital models are created by combining hundreds or even thousands of photographs taken from different angles.
The result is a highly realistic virtual version of the location that investigators can study long after the physical scene has been cleared.
The new research takes this idea much further. The team is developing AI software that can automatically recognize and label objects inside the virtual crime scene. Investigators could simply ask questions such as, “Where is the red jacket?”, “How many knives are in the room?” or “Show me every sharp object.”
Instead of searching through hundreds of photos by hand, the AI would quickly identify the relevant items and highlight them in the digital model.
The researchers hope this technology will eventually combine many different types of evidence into one interactive virtual environment. Physical objects, their locations and other evidence could all be connected in a way that helps investigators better understand what happened and how events may have unfolded.
The team has already made progress in speeding up the creation of these digital crime scenes. Since documenting a crime scene often produces enormous numbers of photographs, processing all of them can take a long time.
The researchers have developed a method that automatically detects pictures that add little or no new information. By removing unnecessary images from the process, they can create the 3D models more efficiently without losing important details.
The next stage of the project focuses on understanding how objects relate to one another in space. For example, the system could calculate whether two people standing in different parts of a room could see each other or determine exactly which areas were visible from a particular position. These kinds of analyses could provide valuable information during criminal investigations.
The project is a close partnership between scientists and police investigators. Researchers at TUM develop the AI and image analysis methods, while experts from the Bavarian State Criminal Police Office help ensure the technology meets the practical needs of real investigations and can be used in everyday police work.
The first tools developed through the project are now being introduced into the workflows of the Bavarian and Hessian State Criminal Police Offices. Looking ahead, the researchers believe the technology could become even more accessible.
One day, police officers arriving at a crime scene may be able to use an ordinary smartphone to capture images that are automatically turned into an intelligent 3D crime scene.
As the technology continues to improve, these interactive digital twins could help investigators examine evidence more quickly, better understand crime scenes and solve cases more efficiently.
