filmov
tv
Car Counting with YOLO and OpenCV in C++
Показать описание
This is the result of a university project that I was working on using YOLOv3, OpenCV 3.4.9 and C++ to detect and count vehicles in a given video footage.
At the end of the video the display in the top right corner tells us that at 416x416 input frame resolution the program detects 224/275 cars (81% accuracy) at an average of 178.2 milliseconds of processing time (5.6 FPS).
At the end of the video the display in the top right corner tells us that at 416x416 input frame resolution the program detects 224/275 cars (81% accuracy) at an average of 178.2 milliseconds of processing time (5.6 FPS).