Underground roadway accumulated water detection and identification method

Abstract

The invention relates to an underground roadway accumulated water detection and identification method, which comprises the following steps of: performing detection network model training based on a data sample set, including performing CNN model pre-training of underground roadway accumulated water detection, performing RPN model training of underground roadway accumulated waterdetection and performing SVM classifier training; acquiring a be-detected image of the underground roadway based on a mobile camera; and based on a detection network model obtained by the CNN model pre-training, the RPN model training and the SVM classifier training, carrying out detection identification on the to-be-detected image to obtain an underground roadway accumulated water condition. The image of the underground roadway is detected and analyzed based on the trained detection network model to determine whether the accumulated water exists, and the underground roadway accumulated water detection problem in complex environments such as darkness, no light, water surface reflection and the like can be effectively solved.

Description

technical field [0001] The invention relates to the field of inspection of mine roadways, in particular to a method for detecting and identifying accumulated water in underground roadways. Background technique [0002] Daily inspection of dangerous factors such as dangerous gas and water accumulated in the mine roadway is a vital link to ensure mine production and personnel safety. The water in the roadway will make the air humid, which will easily cause corrosion of various electrical production equipment in the mine, resulting in short-circuit faults; at the same time, the roadway has a complex environment such as narrow space, weak light, and high dust. Therefore, how to effectively and automatically detect and identify water in the roadway has a Important research value and application value. [0003] At present, most mine water detection methods are implemented based on visual sensors. Because they are easily limited by light, additional light sources are usually requi...

AI Technical Summary

Problems solved by technology

The water in the roadway will make the air humid, which will easily cause corrosion of various electrical production equipment in the mine, resulting in short-circuit faults; at the same time, the roadway has a complex environment such as narrow space, weak light, and high dust. Therefore, how to effectively and automatically detect and identify water in the roadway has a Important research value and application value

[0003] At present, most mine water detection methods are implemented based on visual sensors. Because they are easily limited by light, additional light sources are usually required for supplementary light. This not only increases the implementation cost, but also makes it difficult to extract the edge of the water surface due to reflections on the water surface. information, thereby affecting the test results

There are also some detection methods that need to preset corresponding sensors at fixed locations, which are not suitable for long-distance water accumulation detection in roadways

Images