|
|
Recognition system of occlusion on segment surface of subway tunnel |
HUANG Yuanyuan, HAO Peng, SUN Yi
|
Shanghai Investigation, Design & Research Institution Co., Ltd., Shanghai 200434, China |
|
|
Abstract The subway tunnel disaster recognition based on image processing is affected by the occlusion on the surface of the subway tunnel segment, which leads to inaccurate disaster feature recognition. In order to reduce the influence of occlusion on disaster recognition, according to the characteristics of various occlusions on the surface of the subway tunnel, the Mean-shift target tracking algorithm was used to determine the center position of the occlusion and then determine its area. The cascade classifier was used to quickly identify the occlusion and mark all the occlusions. This method could quickly identify and mark the occluded objects in the subway tunnel segment image. When there was no accurate identification, the algorithm would intelligently prompt and record the unidentified image for subsequent artificial auxiliary marking. Experiments showed that this method could quickly batch process subway tunnel segment image. In the disaster identification of subway tunnel, the influence of shelter area can be avoided, and it has certain engineering practicability.
|
Received: 28 May 2021
Published: 20 December 2021
|
|
|
|
|
|
|