Current Status and Future Directions of Deep Learning Applications for Safety Management in Construction

Abstract

The application of deep learning (DL) for solving construction safety issues has achieved remarkable results in recent years that are superior to traditional methods. However, there is limited literature examining the links between DL and safety management and highlighting the contributions of DL studies in practice. Thus, this study aims to synthesize the current status of DL studies on construction safety and outline practical challenges and future opportunities. A total of 66 influential construction safety articles were analyzed from a technical aspect, such as convolutional neural networks, recurrent neural networks, and general neural networks. In the context of safety management, three main research directions were identified: utilizing DL for behaviors, physical conditions, and management issues. Overall, applying DL can resolve important safety challenges with high reliability; therein the CNN-based method and behaviors were the most applied directions with percentages of 75% and 67%, respectively. Based on the review findings, three future opportunities aiming to address the corresponding limitations were proposed: expanding a comprehensive dataset, improving technical restrictions due to occlusions, and identifying individuals who performed unsafe behaviors. This review thus may allow the identification of key areas and future directions where further research efforts need to be made with priority.

Research on Dynamic Monitoring and Early Warning of the High-Rise Building Machine during the Climbing Stage

“…Various types of sensors (e.g., cameras and laser ranging sensors) can be added to the acquisition and analysis system to further improve accuracy, and data fitting can be used to identify the spatial relationship between two monitoring points. In addition, the deep learning approach can be considered in future studies on construction equipment safety management [36], such as predicting the motion attitude of climbing equipment by accumulating a large amount of climbing data.…”

An accurate fake news detection approach based on a Levy flight honey badger optimized convolutional neural network model

“…At the prior moment, the status degree information is ignored by determining the reset gate is $$$ {UP}_k^{(t)} $$$. The bias matrix is B and the weight matrix WE 37 . Where $$$ {O}_k^{\left(T-1\right)} $$$ is the output based on a hidden state and at step T , the reset information is represented as $$$ {\tilde{O}}_k^{(T)} $$$.…”

Análise de técnicas de data augmentation para aperfeiçoamento da detecção de sistemas de guarda-corpo e rodapés em canteiros de obras com inteligência artificial

“…Os autores destacam a necessidade de processos automatizados envolvendo outras tecnologias digitais, principalmente a visão computacional. Observa-se na literatura o aumento das aplicações de visão computacional em inspeções de segurança em canteiro de obras nos últimos anos (AKINOSHO et al, 2020;PHAM et al, 2021;COSTA, 2022). São identificadas implementações principalmente de técnicas de aprendizado de máquina (AM) e, mais especificamente, de aprendizado profundo (AP) (campos da inteligência artificial) na detecção de equipamentos de proteção individual e de SPCQ, entre outras aplicações (AKINOSHO et al, 2020;PHAM et al, 2021;COSTA, 2022).…”

“…Observa-se na literatura o aumento das aplicações de visão computacional em inspeções de segurança em canteiro de obras nos últimos anos (AKINOSHO et al, 2020;PHAM et al, 2021;COSTA, 2022). São identificadas implementações principalmente de técnicas de aprendizado de máquina (AM) e, mais especificamente, de aprendizado profundo (AP) (campos da inteligência artificial) na detecção de equipamentos de proteção individual e de SPCQ, entre outras aplicações (AKINOSHO et al, 2020;PHAM et al, 2021;COSTA, 2022). Ottoni et al (2023) pontuam que um desafio para a aplicação de técnicas de AP no setor da construção consiste, muitas vezes, na disponibilidade de pequenos bancos de dados para treinamento desses algoritmos de detecção.…”