Flame detection system of graphic processing based on MATAB GUI

edit  

  Change a video to get data results

The popularity of network and multimedia technology , Flame detection in infrared images The scientific research value of identification technology has been fully demonstrated in various fields . Disaster prevention and relief , Flame detection in infrared images Recognition technology can quickly identify and lock the earthquake covered by dust 、 Trapped people in scenes such as mine disaster and fire smoke , Buy time for effective and timely rescue work , Save lives to the greatest extent and provide security ; On the military side , The application of this technology in weapon aiming system , Be able to quickly find and determine the cover or the enemy in the dark , The coordinated combat ability of the army in bad weather and at night is greatly improved ; Public services , Relevant technologies are applied to the safety monitoring of public and special scenes 、 Traffic monitoring 、 Hunting and personnel rescue have played a key role .

In this paper, infrared image processing with matlab Of rgb Function circle the human body , Then use the projection method ： Projection , Vertical projection , Horizontal projection , Research human body station 、 lie These two kinds of static and walk 、 Fall these two dynamics , Then the human identification flame detection is obtained Technical analysis of , With a view to the current infrared image flame detection Identify relevant research to provide reference suggestions .

Face flame detection Recognition technology has broad application prospects and important academic research significance , Many domestic and foreign researchers and relevant businesses have shown great interest in scientific research in this field . At the same time, it drives the projection based image 、 infrared 、 Research on information related fields of visible light image technology .

    National Institute of information and automation, France (INRIA ) Bill Triggs[4] The research team started from 2000 Since, he has been engaged in visual analysis and Research on human motion posture , Its members C. Sminchiseseu stay 2001-2004 In, a lot of work was done in human motion estimation with monocular camera , Most of the research done belongs to the production model ;2004 After, one of the main tasks is to detect human motion and analyze motion posture , Robust description of human body shape , Regression of motion data and shape parameters by machine learning , So as to reconstruct three-dimensional human motion [5].

Brown University, USA Michael J Black The leading vision group is dedicated to model-based human motion analysis , The main content is to extract human body region features from monocular or multi camera video images , Including contour edges , Compare the similarity between the obtained features and the projection of the model , Then the method based on annealing particle filter or graph model is used for tracking [6]. Besides , Brown University Leonid Sigal The professor has not only made excellent achievements in attitude estimation algorithm , At the same time, it also provides a public flame detection for researchers database HumanEva[7]. data The library provides a series of motion videos taken by a certain number of experimenters and the true values of three-dimensional posture of limbs given by the corresponding motion capture devices , So that researchers can easily obtain experimental data . At the same time, the database can also provide a unified platform for the comparison between different tracking algorithms proposed by different research groups .

The Institute of automation of the Chinese Academy of Sciences has made good research achievements in gait recognition and visual monitoring , Based on two-dimensional video image , Its system can identify the incoming identity from a long distance [8]. The image computing group of Microsoft Research Institute has proposed a texture based statistical method to describe human motion . Through the capture of human motion , get data , After three-dimensional reconstruction, the basic dance movements are generated , And generate new human actions according to the model , It can be used to choreograph and change detailed movements , This study only focuses on the motion analysis of the skeleton , Non rigid deformation of limbs is not involved yet . The Joint Laboratory of Zhejiang University and Microsoft visual perception has invested a lot in animation acquisition 、 Human movement is restricted by access devices , This paper presents a video based human animation technology , A video animation system based on dual cameras is developed , The system should stick markers on human joints , 3D structure and motion acquisition similar to stereo vision .

RGB( Red, green and blue ) It's a space defined by the color recognized by the human eye , Can represent most colors [9]. But in scientific research, it is generally not used RGB Color space , Because its details are difficult to adjust digitally . It will tone , brightness , Saturation three quantities put together to express , It's hard to separate . It is the most general hardware oriented color model . This model is used in color monitor and a large class of color video camera . stay RGB In the color space , The chromaticity of light depends on R, G, B, Only two chromaticity coordinates are independent of each other . This article is based on RGB Function formula for image signal conversion , For different projection methods , Calculate the flame detection Features identified .

Source code see ： be based on matabGUI Graphics processing flame detection system -Matlab Document resources -CSDN download

If there is any infringement , Please contact the author for deletion