Fire early warning method and system based on semantic segmentation and recognition
A technology of fire early warning and semantic segmentation, which is applied to fire alarms that rely on smoke/gas effects, character and pattern recognition, neural learning methods, etc., can solve the problems of poor detection of small target fire areas and achieve good segmentation effects The effect of enhancing accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0036] Such as figure 1 As shown, the present embodiment provides a technical solution: a fire early warning method based on semantic segmentation and recognition, comprising the following steps:
[0037] Step 1: First send the image to be tested to the segmentation network to obtain the segmentation result of the fire area;
[0038] Step 2: Combine the segmentation results output by the segmentation network with the image to be tested to obtain images of the fire and the area around the fire;
[0039] The output of the segmentation network is a probability map. After binarization, a binary map is obtained, and the connected area analysis is performed on the image. Each connected area is a suspected fire area, and the connected area with a smaller area is filtered out. , find the minimum circumscribed horizontal rectangle for the remaining connected areas, and then map this rectangle to the image to be tested to obtain the suspected fire area image; in short, this step obtain...
Embodiment 2
[0050] In this embodiment, the image to be tested is first input into the segmentation network to obtain the segmentation result of the fire area; as Figure 4 shown; then according to the segmentation results, combined with the original image to be tested, the fire and its surrounding areas in the image are obtained. In practical applications, according to the size of the fire area, it is enlarged by 1.3 times as the final fire area. Such as Figure 5 shown. Input the pictures of the fire and its surrounding area into the recognition network for identification, if it is a fire, it will give an early warning, if it is not a fire, it will not give an early warning. non-fire areas such as Image 6 shown.
[0051] In summary, the fire early warning method based on semantic segmentation and recognition in the above embodiment uses a deep learning network to segment the fire area, which is better than the traditional manual method; it makes full use of the prior knowledge that s...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More