Poyang Lake Ecological Economic Zone has large forest area, so it's very important to construct prevention forest fire disaster system. This paper presents an algorithm for prevention forest fire disaster based on digital image processing technology. The algorithm distinguishes the realtime forest video by smoke and fire. To determine whether there are some suspicious area in the image in the spatial domain by judging the color properties of smoke and fire through Clustering Algorithm. If it detects any suspicious circumstances, then fixes ccd and detects the suspicious areas in the time domain. In this step, firstly get the initial detect results by wavelet decomposition , then use the k-means clustering algorithm for the spread detection of smoke. Experimental results show that the algorithm is ideal for the experimental video. It alarms before the fire disaster occurs to avoid major fire disaster, which protects the forest resources in the Poyang Lake Ecological Economic Zone.
Unmanned Anti-disaster System R^|^amp;D Project （NEDO）
