The Study Of Intelligent Monitoring System Based On Qt And OpenCV

Monitoring system is always playing an important role in human life and work. In recent years, along with the rapid development of image processing technology and computer processing ability, the research and implementation of intelligent monitoring system has become one of the development directions of monitoring system. In the study of intelligent monitoring system, to detect moving object, namely real-time extracting and matching the feature of color or background from each frame in the video image, was the most basic application and of the most important function in intelligent video monitoring system. So the research of this technology becomes an important subject in the field.The purpose of this paper is to design a multithreaded indoor environment intelligent monitoring system with more cameras. The main function of the monitoring system as follow:The system can be controlled by thread at the speed of 30 frames/s acquisition four cameras video images at the same time; Capture video can be segmented in the process of preservation to facilitate staff in daily management and issue verification has a simple and practical human-machine interaction interface, the four cameras can be opened and closed, and has buttons of intelligent model option, choose to implement according to the corresponding buttons or suspicious objects moving target detection and tracking; according to the corresponding button, we can detect and track the moving target; For indoor environment with color characteristics of things or situations can be real-time detected and timely alerts or messages to remind units or personnel.This topic mainly involves the relevant contents of monitor system as follows: Video images ingestion based on multi-camera multithreaded,camera calibration, image preprocessing, image edge detection, the method of moving object detection and tracking based on optical flow, target detection technology based on color feature, Qt and Open CV configuration methods and procedures under the dynamic link library encapsulation and so on.Through the study of this article, the main implementation of the content as follows:(1) Using the camera calibration toolbox in Matlab to complete the camera calibration based on the analysis of the geometry of the pinhole camera model. The feature points were extracted from 28 images through the corresponding relationship of space coordinate and image coordinate. The camera’s internal parameters and external parameters can be determined with the camera calibration.(2) Writing multiple cameras video image extraction procedure with Qt and Open CV vision libraries as software platform, and by using the correlation function of Open CV to finish image pre-processing, including color image gray, gray image filtering, image edge detection and comparative analysis of various algorithms, ultimately determines the image pretreatment process, and obtained good treatment results.(3) Using optical flow method to achieve the moving object detection and tracking. The system can judge the movement direction of the moving objects indoor environment, and determine its behavior intention, anticipation suspicious personnel and alarm.(4) After the space transformation of color image, the system can realize the bright colors such as red(fires), blue(uniform) using Cam Shift algorithm based on color in the indoor environment with color feature detection and tracking of things.(5) By writing monitoring system of the man-machine interface, realize the real-time control of the four cameras, and finally complete the intelligent monitoring system software package, which can be used as a file on a computer or other device to use directly.