Method and system for automatically opening tunnel fireproof door by inspection robot in dark and humid environment

Abstract

The invention provides a method for automatically opening a tunnel fireproof door by an inspection robot. The method comprises the following steps: acquiring image data or video data on a motion track of the robot in a dark and humid environment; according to the acquired data, utilizing a preset neural network to extract shallow layer features and deep layer features of the network, performing deconvolution on multiple layers of convolution features, fusing deconvolution results, and generating a saliency map of the target image; target recognition is carried out on the obtained saliency map, a target classification probability is obtained, and then a fireproof door recognition result is obtained; when the fireproof door is recognized, a door leaf state query instruction is generated and sent to the door leaf control terminal, and the door leaf control terminal obtains the current state of the door leaf according to the door leaf state query instruction and controls opening or closing of the door leaf according to the current state of the door leaf; the fire door identification precision is improved, the problem that the inspection robot is low in inspection efficiency due to the fact that the fire door is not opened or closed in time is solved, and it is guaranteed that the inspection robot safely passes through the fire door.

Description

technical field [0001] The invention relates to the technical field of tunnel inspection, in particular to a method and system for automatically opening a tunnel fire door by an inspection robot in a dark and humid environment. Background technique [0002] The statements in this section merely provide background related to the present disclosure and do not necessarily constitute prior art. [0003] In the long and narrow buildings such as urban comprehensive pipe gallery and power cable tunnel, there are a large number of pipelines and equipment, among which the lighting conditions are poor, the air humidity is relatively high, and the harmful gas content is high. The operation risk is high, and the introduction of inspection robots can significantly improve work efficiency and ensure operation safety. Robot inspection has been accepted by more and more operation and maintenance units. However, in order to meet the requirements of fire zoning, a normally closed fire door ...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the deficiencies of the prior art, the present invention provides a method and system for automatic opening of tunnel fire doors by inspection robots in dark and humid environments. The improved fire door identification algorithm is adopted to improve the identification of fire doors in poor lighting conditions and humid and fuzzy environments. Accuracy, which solves the problem of low inspection efficiency caused by the untimely opening and closing of fire doors by inspection robots, and ensures that inspection robots pass through fire doors safely

Images