Banca de DEFESA: ALISSON DE SOUZA DA SILVA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : ALISSON DE SOUZA DA SILVA
DATE: 22/03/2023
TIME: 15:00
LOCAL: Sala PPEC
TITLE:

METHOD FOR AUTOMATED RECOGNITION OF PATHOLOGIES  ON EXECUTION OF FACADES USING UAS AND MACHINE LEARNING

KEY WORDS:

 Automated inspection. Digital technologies. Drone. Artificial intelligence. Quality control.

PAGES: 170
BIG AREA: Engenharias
AREA: Engenharia Civil
SUMMARY:

The facade is one of the main elements of a building and has functions that range from protecting the internal structure of the building to ensuring the health and well-being of its users. The emergence of pathological manifestations in these elements can compromise their performance; therefore, periodic inspection and maintenance activities are necessary. However, inspections are often not carried out or neglected as they are time-consuming, expensive, and unsafe. Given this, digital technologies, such as Drones and Artificial Intelligence-AI, can minimize such limitations. Therefore, the main objective of this work is to propose an automated recognition method to identify pathological manifestations in concrete facades using drones and AI to support decision-making for the quality control of works. For this, the Design Science Research-DSR research strategy was adopted, and the study was developed according to the following steps: (1) Awareness through the investigation of the theoretical problem through a Systematic Review of Literature-RSL, identifying the research gaps, and through practical problem investigation in the field; (2) Suggestion of the artifact through an exploratory study, in which drones were used to acquire images and AI algorithms for digital image processing; (3) Development of the artifact with the accomplishment of two case studies, implementing the method of automated recognition of pathological manifestations through a data acquisition protocol, data processing protocol, and incorporation of information from these activities in the process of control of the quality of work; (4) Evaluation of the artifact through the constructs Transparency, Efficiency, Applicability, and Utility through the variables for each of these constructs; and (5) Conclusion with the formalization of the proposed method, final considerations about the study, and communication of the results. As preliminary results, data acquisition and processing protocols were implemented in Case Study B and are being improved and refined in Case Study C. To date, 2040 images with pathological manifestations have already been collected from the acquisition of images. Regarding the processing of these images, four models were created in Custom Vision and are being improved with new images from visits and images generated from on-site photos using the Data Augmentation technique. The information from the inspections carried out during the visits is being implemented in Case Study B and C through the delivery of reports forming a PDCA cycle

COMMITTEE MEMBERS:
Presidente - 2618348 - DAYANA BASTOS COSTA
Interno - 1522482 - FRANCISCO GABRIEL SANTOS SILVA
Externo à Instituição - Fabiano Rogerio Corre?a
Interno - 1315606 - REYMARD SAVIO SAMPAIO DE MELO
Externa à Instituição - ROSENEIA RODRIGUES SANTOS DE MELO