Computer vision applications in construction: Current state, opportunities &amp; challenges

Paneru, Suman; Jeelani, Idris

doi:10.1016/j.autcon.2021.103940

Cited by 155 publications

(42 citation statements)

References 89 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…CV is a broad term that refers to various techniques for extracting and processing visual data from images and videos to make inferences. Some of these techniques that are important to construction management tasks are: 1) 3D scene reconstruction, 2) Image and object classification, 3) Object recognition, 4) Object tracking, 5) Segmentation, and 6) Action recognition (Paneru and Jeelani, 2021). The tasks and related challenges are furtherly explained in Table 2 below.…”

Section: Computer Vision and Natural Language Processingmentioning

confidence: 99%

Artificial intelligence in construction asset management: a review of present status, challenges and future opportunities

Rampini¹,

Cecconi²

2022

ITcon

View full text Add to dashboard Cite

The built environment is responsible for roughly 40% of global greenhouse emissions, making the sector a crucial factor for climate change and sustainability. Meanwhile, other sectors (like manufacturing) adopted Artificial Intelligence (AI) to solve complex, non-linear problems to reduce waste, inefficiency, and pollution. Therefore, many research efforts in the Architecture, Engineering, and Construction community have recently tried introducing AI into building asset management (AM) processes. Since AM encompasses a broad set of disciplines, an overview of several AI applications, current research gaps, and trends is needed. In this context, this study conducted the first state-of-the-art research on AI for building asset management. A total of 578 papers were analyzed with bibliometric tools to identify prominent institutions, topics, and journals. The quantitative analysis helped determine the most researched areas of AM and which AI techniques are applied. The areas were furtherly investigated by reading in-depth the 83 most relevant studies selected by screening the articles’ abstracts identified in the bibliometric analysis. The results reveal many applications for Energy Management, Condition assessment, Risk management, and Project management areas. Finally, the literature review identified three main trends that can be a reference point for future studies made by practitioners or researchers: Digital Twin, Generative Adversarial Networks (with synthetic images) for data augmentation, and Deep Reinforcement Learning.

show abstract

Section: Computer Vision and Natural Language Processingmentioning

confidence: 99%

Artificial intelligence in construction asset management: a review of present status, challenges and future opportunities

Rampini¹,

Cecconi²

2022

ITcon

View full text Add to dashboard Cite

show abstract

“…The state of the art of automatic progress identification using point clouds is a process that has six significant steps: image collection, point cloud generation, registration, point cloud segmentation, point cloud semantic recognition, and progress reasoning [15]. Despite this, [15] and [16] state that this process still has limitations, such as the challenge of performing automatic alignment (registration) between the BIM model and the point cloud, the presence of noise and unwanted objects in the point cloud, and the high computational cost to perform the semantic recognition of objects in the point cloud to infer the status of activities.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Furthermore, data privacy remains a concern for workers whose activities get captured by these visual sensors. The question of who owns data and how the data can be used remains unclear [16].…”

Section: Literature Reviewmentioning

confidence: 99%

Use of BIM and visual data collected by UAS and 360° camera for construction progress monitoring

Barbosa

Costa

2022

IOP Conf. Ser.: Earth Environ. Sci.

View full text Add to dashboard Cite

Despite progress monitoring is an essential practice for achieving the success of construction, traditional monitoring methods based on manual information gathered through visual inspections are error-prone, depending on the experience of those who carry them out. Furthermore, most studies of progress monitoring using digital technologies focus on activities carried out outdoors, limiting the application of these methods in residential construction sites, which have several indoor activities. This study proposes a method for outdoor and indoor visual monitoring of construction progress using Building Information Modeling (BIM), 360° camera, and photogrammetry aided by an Unmanned Aerial System (UAS). For this purpose, exploratory case studies were carried out. The first exploratory study aimed to understand data collection and processing operationalization using the proposed technologies. Then, these technologies were used and evaluated to monitor progress in a second exploratory case study, enabling the development of a proposed method for using visual data collected by UAS and 360° camera integrated to BIM for progress monitoring. The status of the external area of the construction site was represented by point clouds generated through images collected by UAS. For monitoring inside the buildings, a 360° camera attached to the safety helmet was used. The results include evaluating the use of a 360° camera to monitor the internal progress of works, presenting its strengths, limitations, and use recommendations. In addition, the results also include the proposal of a method for visual progress monitoring of indoor and outdoor activities using BIM, UAS, and 360° cameras.

show abstract

“…Computer Vision is concern with enabling computer to derive meaningful information from digital images, video and other visual input. Computer vision has variety of applications such as; Security [1,2] manufacturing [3], construction [4,5] etc. [6] in their book provides an overview of computer vision along with its applications.…”

Section: Introductionmentioning

confidence: 99%

A Comparative Study of Eigenface and Fisherface Algorithms Based on OpenCV and Sci-kit Libraries Implementations

Aliyu¹,

Bomoi²,

Maishanu³

2022

IJIEEB

View full text Add to dashboard Cite

Facial Recognition is the task of processing an image or video content in order to identify and recognize the faces of individuals. Its area of applications are wide and a lot of research efforts have been invested which led to introduction of techniques/algorithms and programming language libraries for implementation of those techniques. Facial recognition relies heavily on the use of machine learning techniques. Convolutional Neural Network (CNN), a deep learning algorithm has been successfully applied for face recognition task. However, because of its requirements, it may not be applicable in all cases. Where application scenario cannot cope with CNN, it is necessary to resort to other techniques that use traditional Machine Learning (ML) techniques. Previous studies that performed comparison on face recognition algorithms that use traditional ML techniques only disclosed the best algorithm without revealing the best image processing library used. Considering the fact that people now depend on these libraries to build face recognition systems, it is important to empirically show the best library. In this paper an experiment was conducted with aim of assessing the performance of Fisherface and Eigenface algorithms, and that of Scikit-learn and OpenCV libraries. Eigenface and Fisherface algorithms were combined with K-Nearest Neighbors (KNN) and Support Vector Machines (SVM) classifiers respectively. The algorithms were evaluated using LFW dataset, and implemented in two Python libraries for image processing Scikit-learn and OpenCV. This is to enable us determine the best performing technique/algorithm and at the same time the best library, thereby achieving dual aims. Experimental results show that Scikit-learn implementation of Fisherface with KNN recorded the highest F-score of 67.23% while the OpenCV implementation of Eigenface with SVM recorded the lowest F-score of 14.53%. Comparing the algorithms, Fisherface with SVM produced better results than Eigenface with SVM. The same story holds for Fisherface with KNN, and Eigenface with KNN. This suggests that irrespective of classifier, Fisherface outperform Eigenface in terms of accuracy of recognition. Comparing the libraries, Scikit-learn implementations of Fisherface with SVM and Eigenface with SVM, outperform the OpenCV implementation of the same algorithms. This means scikit-learn implementation produces better results than its counterpart, the OpenCV.

show abstract

Computer vision applications in construction: Current state, opportunities & challenges

Cited by 155 publications

References 89 publications

Artificial intelligence in construction asset management: a review of present status, challenges and future opportunities

Artificial intelligence in construction asset management: a review of present status, challenges and future opportunities

Use of BIM and visual data collected by UAS and 360° camera for construction progress monitoring

A Comparative Study of Eigenface and Fisherface Algorithms Based on OpenCV and Sci-kit Libraries Implementations

Contact Info

Product

Resources

About