Smog detection method and device based on video analysis

Abstract

The invention discloses a smog detection method based on video analysis. The smog detection method based on video analysis comprises steps of video image data acquisition, video image preprocessing, forward target extraction, flame static characteristic detection, flame dynamic characteristic detection, flame alarm decision, smog static characteristic detection, smog dynamic characteristic detection and smog alarm decision. The method is advantaged in that smog and flame are simultaneously detected in a combined static and dynamic characteristic mode, smog detection accuracy and stability are improved, operation complexity is further reduced, hardware realization is convenient, and the method can be applied to relatively complex environments.

Description

technical field [0001] The invention belongs to the technical field of intelligent fire monitoring, and in particular relates to a method and device for detecting smoke and fire based on video analysis. Background technique [0002] Fire detection systems have important applications in many fields, such as fire prevention in forests, warehouses, oil fields, and farms. Traditional pyrotechnic detection devices mainly use some sensors to detect pyrotechnics by detecting thermal airflow, solid suspended particles in smoke, etc. These detection devices need to be in close contact with the fire source, which cannot meet the needs of open scenes such as forests and farms. In addition, in the initial stage of fire combustion, smoke is usually released, so smoke detection is also very important for early warning of fire. In view of the above, a video-based smoke detection method is proposed. [0003] The current video-based pyrotechnics detection methods mainly include: using the ...

AI Technical Summary

Problems solved by technology

[0006] In the prior art, the method of judging whether it is a flame based on the morphological characteristics of the flame, such as judging whether it is a flame based on the circularity and sharp corner features, and judging whether it is a smoke based on characteristics such as diffusivity, upward movement, and background blur is likely to cause false positives, because flames and smoke The size and shape of the tree are very unstable, most of the changes are not continuous, and it is easily affected by wind and light, and sometimes it is difficult to meet the above characteristics

[0007] (2) Easy to misreport

[0008] In the prior art, the method of judging whether there are fireworks by using color features alone combined with a small amount of shape features or motion features is prone to allergic false alarms for irregular motion shapes, such as non-rigid objects such as people, leaves, and shadows.

[0009] (3) The amount of calculation is large, and the hardware implementation is difficult

[0010] In the prior art, the method of detecting pyrotechnics based on the accumulation of the main motion direction tends to cause large memory consumption, and the method of detecting pyrotechnics using frequency domain features of wavelet transform requires high calculation accuracy, and hardware implementation has limitations

Images