The new development from XDSOFT's R&D lab is a contour recognition and detection system for surveillance cameras
XDSOFT has introduced a neural network application designed to process frames from photo and video surveillance systems (such as outdoor security cameras) for the purpose of contour detection of specified objects and their classification by type (e.g., car, truck, motorcycle, bicycle, pedestrian, etc.), and recording based on predefined parameters.
How the system works:
-
Object Recognition and Classification: The client connects to the processing server and uploads an image for recognition and further processing. The server processes the file to recognize, count, and classify objects, such as cars, trucks, pedestrians, bicycles, and motorcycles.
-
Contour Detection: The system defines the boundaries of each recognized object, marking its contour to differentiate it from the background.
-
Output Generation: The system can generate different output formats, such as an image with color-filled areas, an image with defined contours, or a file (e.g., XML, JSON) containing details like the number and type of recognized objects, an array of contour coordinates, and other necessary descriptions.
-
Delivery of Processed Results: The processed material is sent back to the client.
The system can be used for real-time object monitoring or for periodic inspections of object presence (e.g., monitoring parked cars, checking the presence of road signs, or other objects during routine inspections).
