Leaked gas cloud cluster detection method based on thermal imaging video

The invention discloses a leaked gas cloud cluster detection method based on a thermal imaging video, and belongs to the field of gas detection. The leakage gas detection model based on the shape features of the gas infrared image is obtained by using the support vector machine and SIFT feature training, automatic detection of the leakage gas is realized, the leakage condition of the gas is displayed in a video form by using the visibility of the leakage gas in an infrared scene, and the method has the advantages of large-range imaging and direct positioning of a leakage source. The SIFT features are used as feature data of leaked gas to be input into the support vector machine, the advantages of small feature data size, low model complexity and short training and detection time of the SIFT features are fully utilized, and the detection accuracy of the leaked gas is improved by reducing the complexity of a leaked gas detection support vector machine model. And the time cost and the storage space cost of the l