DESIGN AND APPLICATION OF FOREST FIRE DETECTION SYSTEM BASED ON IMAGE RECOGNITION TECHNOLOGY

摘要

In recent years, image recognition technology and wireless digital microwave communication tech-nology have developed rapidly and have been widely used in many fields. In order to realize the remote real-time monitoring of the environmental status of the forest, this paper combines image recognition technology, H.264 video compression technology and wireless digital microwave communication tech-nology to design a forest fire monitoring system. The monitoring system can collect the video images of the forest in real time and send them to the remote monitoring center. The monitoring center realizes the real-time monitoring function of the forest by identifving the video images for fire. Fire image recognition uses fire detection algorithms based on RGB, YCbCr and HSIcolor spaces. In this study, we convert the RGB image to YCbCr and HSI color space and extract the grayscale image of each chan-nel of the three-color spaces,use PCA dimensional-ity reduction method to analyze the specificity of each channel, and establish the ROC curve to obtain the identification criterion threshold. Through exper-iments, this monitoring system can complete the col-lection and transmission of forest video images, and perform reliable analysis and processing of image data, thereby realizing real-time monitoring of forest environmental conditions.