Study On Methods Of Infrared Forest Fire Image Target Detection

Forest fires is a major reason for the global forest resource diminishes, which not only caused the huge loss of resource, but also had a great influence on the global climate. How to use the infrared image information to quickly detect the fire point of the targets has become one of the hot researches at home and abroad. Therefore, this article is for the forest fire target detection method based on infrared technology.Research work carried out in this paper are:Firstly, further study the small target detection methods of infrared forest fires,often appearing the problems of detection error or undetected,so using the neighborhood contrast can enhance small target and suppression the background,using multi-scale template matching to determine the size of small targets.Experiments has shown the algorithm could improve the detection accuracy and the signal to noise ratio of the image.Secondly, due to the low SNR of collected infrared forest fire image, existing algorithms is difficult to improve the SNR and extracted the target at the same time.In view of this, the paper uses the improved Otsu, tectonic penalty function, let the problem of Otsu transformed into seeking a function’s optimal solution with the constraints.Experiments show that the algorithm is a more robust object extraction algorithm.Finally, in order to overcome imperfect information of the single image sensor,further study several fusion methods advantages and disadvantages,it always let the fused image have a fuzzy edges and the detail is not obvious, so this article is using NSCT and improved PCNN for image fusion.Utilizing NSCT decomposition gets high and low frequency coefficients, with the region’s largest energy deal with the low-frequency coefficients, high-frequency coefficients integration is to choice the maximum of local variance、local sharpness and spatial frequency, finally get the fused image by NSCT inverse transformation. Fusion results are ideal, the image details get a wealth of abundant, the entropy and clarity,etc. has been improved.