Visual profiler accelerates target processing from aerial, satellite imagery

3841
An image showing multi-type processing of the system, showing a specific vehicle type among other targets spotted on the roofs. Photo: VideoInform (from screen).
Visual Profiler, developed by Video Inform and operationally deployed for automatic analysis of aerialimages can scan large image databases, real-time or historic, to detect targets of interest. This example shows the systems spotting pick up truck shapes vehicles using automatic detection. The system can further recognize even finer details, such as specific distinguishable details such as color, make, and unique, distinctive characteristics through automtaic-processing
Visual Profiler, developed by Video Inform and operationally deployed for automatic analysis of aerial images can scan large image databases, real-time or historic, to detect targets of interest. This example shows the systems spotting pick up truck shapes vehicles using automatic detection. The system can further recognize even finer details, such as specific distinguishable details such as color, make, and unique, distinctive characteristics through automatic processing. Photo: Video Inform.

Video Inform is introducing an advanced target detection and acquisition capability at the AUSA 2014 event, unveiling the  ‘Visual Profiler’, a cutting-edge cognitive vision technology extracting intelligence information from aerial or satellite imagery. The system can be used as a stand-alone solution or as an add-on to an existing Geographical Information Systems (GIS) and imagery management system.

The Visual Profiler is empowered by a proprietary visual search engine developed by the company, providing automatic detection and profiling of targets from aerial or satellite images in real time. This capability enables the rapid delivery of intelligence information to forces in the field, thus shortening the sensor-to-shooter cycle. Unlimited by the number of target-object definition and profiling objects, the Visual Profiler can be trained to recognize specific vehicles, infrastructure and foundations. A unique interactive and intuitive user feedback mechanism continuously trains and improves the level of precision, further adapting the system for the user’s needs.

“Our Visual Profiler represents a novel approach to image understanding.” said According to Yoram Sagher, CEO, “Based on a unique cognitive vision and profiling methodology, we have tried to imitate the human object recognition process, and to achieve breakthrough performance. The solution provides extremely rapid results,enabling the delivery of intelligence data to the battlefield in real time, thus impacting the outcome while the battle is actually underway.” According to Sagher, the system has been adopted by a leading Air Force, and has received high praise.”

An image showing multi-type processing of the system, showing a specific vehicle type among other targets spotted on the roofs. Photo: VideoInform (from screen).
The system has spotted a water tank on a pickup tuck, ready to transport. Although the typical target (water tank) is mounted on roofs, the system can be taught to detect those targets even in non standard situations, thus highlighting anomalies to analysts. Photo: VideoInform (from screen).
Multiple vehicles of a specific type detected by the system in a parking lot. Video Inform photo, from display
The combination of two target types is demonstrated here, indicating specific vehicle types and their cargo.