Abstract: Three kinds of algorithm were compared for identifying forest fire through infrared thermal imager,the testing method of configuring the algorithm parameters was put forward,and an experiment was made to test this method in Zhuhai.The experiment showed that infrared thermal imager always has higher ability to identify the fire sparks at night than in the daytime.The result of comparing three algorithms that threshold decision method has relatively higher sensitivity but lower anti-interference ability,neighborhood threshold decision method has higher ability of anti-interference but shorter monitoring distance and the average comparison decision method has highest ability of anti-interference but shortest monitoring distance.

Key words: infrared thermal imager, algorithm for identifying forest fire, fire monitoring