Campus Fire Recognition Based on Video Image Processing

The school is a personnel intensive place, and it is easy to cause a fire. Through the video image processing method, the campus fire can be recognized effectively. Traditional campus fire recognition methods based on video image analysis don´t analyze the color, texture and other low-level visual feature information of specific campus fire image. The cross validation performance is not good, and the campus fire image scene recognition results are often unstable, the recognition accuracy is not high. An improved campus fire image recognition and data mining method is proposed based on vision recognition function. The Hough transform is taken to transform the curve or linear of campus fire image into a point in the transformation region. Gray processing and binaryzation processing are taken for the campus fire image. The normalization processing is taken for filtering the background interference factor. The refined video surveillance digital image is obtained, and the visual features acquisition process of campus fire image is analyzed. The small scale Gabor filter is used to collect the visual features of campus fire video monitoring images. According to the different training samples and correlation coefficients, the campus fire image recognition and feature data mining are completed. Simulation results show that the algorithm has better efficiency, it can recognize the campus fire accurately, and give the real time alarming, it has good application value in practice.