Font Size: a A A

Research On Crack Identification And Safety Evaluation Of Masonry Structures Based On Deep Learning

Posted on:2021-11-24Degree:MasterType:Thesis
Country:ChinaCandidate:H Y ZhangFull Text:PDF
GTID:2512306512979479Subject:Civil engineering
Abstract/Summary:PDF Full Text Request
The first sign of structural problems in a house is the cracks in its structural members.This is a fairly common phenomenon and a problem that has long plagued engineering and technical personnel.For structural health and safety monitoring of masonry structures,the primary task is to investigate and study the cracks generated.The main research work of this article is as follows:(1)Use Mask-RCNN to train a large number of crack images,mark the precise location of the crack and the shape trend,verify the recognition accuracy after the training,and compare the image recognition effect with the traditional morphological image processing technology.(2)Extract the mask coordinates identified by Mask-RCNN in python and binarize the crack image.Each crack image contains a target map,which is used as a reference for crack size measurement,thereby measuring the length and width of the crack.In addition,an algorithm check is also made on the shape and direction of the crack.(3)Analyze the causes of cracks in masonry structure houses,consider settlement cracks,temperature cracks,and structural cracks as the main cracks of masonry structure;take the specific location of cracks in the house according to 6 levels And the letter number;according to the requirements of the specification,the judgment conditions of the dangerous components of the masonry structure are optimized into mathematical formulas.(4)Establish a cloud platform website so that each user can upload crack pictures to the crack website.The cloud platform can identify and measure the uploaded crack pictures;a new user questionnaire is proposed for systematic investigation Information on the use history of the building,surrounding environment,temperature and humidity changes,etc.(5)Combining the information of crack length,width,cause,location,shape,etc.,combined with national regulations and the experience of structural safety testing personnel,the health status of masonry structure houses is identified based on crack information.The security evaluation of the masonry structure house in a residential area of Kunshan City was conducted to verify the practicability and efficiency of the method proposed in this paper.
Keywords/Search Tags:Masonry structure, crack, Mask-RCNN, python, structural health identification, crack website
PDF Full Text Request
Related items