Wildfire detection method based on image intelligent analysis

Abstract

The invention discloses a mountain fire detection method based on intelligent image analysis, which includes a training stage and a test stage. The superpixels of the image are obtained by using the SLIC method. The method can quickly cluster the pixels of the image, and can perform well After fitting the edge of the target accurately, after obtaining the superpixel block, its color and texture features are proposed, and the image is classified according to these features. The fire detection method based on texture and color proposed by the present invention can quickly and effectively detect the location of wildfires, and can achieve a more accurate classification result. The superpixels generated by SLIC used in this method can fit the edge of the target well, and the segmentation effect is ideal. The features used to describe the superpixel region are simple and effective, with low complexity, which can meet the real-time requirements of wildfire detection.

Description

technical field [0001] The invention relates to a mountain fire detection method, in particular to a mountain fire detection method based on image intelligent analysis. Background technique [0002] Forest fires refer to fires that occur in the mountains and cause great damage to the forest ecosystem and the safety of human life and property and are difficult to control. They are characterized by strong suddenness, wide coverage, fast spread, and high danger. With global warming, the occurrence of forest fires is becoming more and more frequent. Our country attaches great importance to the prevention and protection of wildfires, aiming to effectively extinguish wildfires in the early stages of their occurrence and prevent their damage to the environment and human life and property. Early fire detection mainly relied on staff patrolling. This method not only consumes a lot of manpower and material resources, but human patrols will inevitably lead to some negligence, and in s...

AI Technical Summary

Problems solved by technology

Although these threshold-based methods are fast and simple, they are very sensitive to the selection of the threshold. Due to the complexity of the forest environment, it is often difficult to select a suitable threshold to distinguish the background from the wildfire.

If the mountain fire area cannot be extracted more accurately, it will affect the subsequent feature extraction, thus increasing the difficulty of fire detection

Images