PROJECT:

In this project, an application was developed that, using a camera, accounts for the number of people who pass, for example, through a door, through facial recognition. In addition, it was also necessary to have the functionality to be able to limit the detection to only a specific region. The application was implemented in C ++, using Qt, as a graphical interface, and the OpenCV library to read images from the camera and face detection, Video Capture and Haar Cascades, respectively. A thread was also used to optimize the entire process.

FEATURES:

• Collection of images in real time from a camera;
• Face detection, even in a single frame;
• Definition of a region of interest;
• Accounting for the number of people.

FILES:

• aplicacao.pro - application project file.
• camera.h and camera.cpp - class camera with the code for the back-end of the developed application and connection with the thread class and mouse_press_event class. It has the function of displaying images, received from the camera, and the results obtained from the thread class.
• camera.ui - interface implemented for the front-end of the final application.
• thread.h and thread.cpp - process and analyze the images. Using the Haar Cascades class, it detects faces in each of the frames received.
• mouse_press_event.h and mouse_press_event.cpp - records the points where the mouse clicked to allow the definition of the region of interest.

RESULTS:

The images on the left are the real-time images captured by the camera while the images on the right are snapshoots of when a face detection occurs.

Last detection performed - green square:

Definition of a region of interest - red square:

Detection performed after defining the region of interest - green square: