@article{jeelani_asadi_ramshankar_han_albert_2021, title={Real-time vision-based worker localization & hazard detection for construction}, volume={121}, ISSN={["1872-7891"]}, DOI={10.1016/j.autcon.2020.103448}, abstractNote={Despite training, construction workers often fail to recognize a significant proportion of hazards in construction environments. Therefore, there is a need for developing technology that assists workers and safety managers in identifying hazards in complex and dynamic construction environments. This study develops a framework for an automated system that detects hazardous conditions and objects in real-time to assist workers and managers. The framework consists of three independent pipelines for localization of workers, semantic segmentation of the visual scene around workers, and detection of static and dynamic hazards. The framework can be used to automate and augment the hazard detection ability of workers and safety managers in construction workplaces. In addition, the framework offers several computing contributions including an improved real-time worker localization method and an efficient architecture for integrating pipelines for entity localization and object detection. A system developed based on the proposed framework as a proof of concept and was tested in indoor and outdoor construction environments. It achieved over 93% accuracy.}, journal={AUTOMATION IN CONSTRUCTION}, author={Jeelani, Idris and Asadi, Khashayar and Ramshankar, Hariharan and Han, Kevin and Albert, Alex}, year={2021}, month={Jan} }
 @inproceedings{boroujeni_han_2017, title={Perspective-Based Image-to-BIM Alignment for Automated Visual Data Collection and Construction Performance Monitoring}, url={http://dx.doi.org/10.1061/9780784480830.022}, DOI={10.1061/9780784480830.022}, abstractNote={In efforts to automate construction performance monitoring, past studies have worked on vision-based registration of image to BIM and 3D point clouds to BIM. The continuous development of simultaneous localization and mapping (SLAM) enabled real-time estimation of locations and orientations of a camera while incrementally reconstructing a 3D scene. However, it localizes a camera to an arbitrary local coordinate system and produces a low-resolution and noisy point cloud that is not suitable for quality assessment of a structure. For the architecture/engineering/construction industry, the better and realistic approach is to localize with respect to building information models (BIMs) in real-time and post-process 3D dense reconstruction. This approach will allow project management teams to better communicate quality and progress using visuals associated with locations shown with BIMs. Moreover, it will automate images-to-BIM and image-based point clouds-to-BIM registration, enhancing past studies that attempt to automate image-based progress detection and quality assessment. On the other hand, the current state-of-the-art method for registering an image-based point cloud to a BIM requires selection of the correspondences. To address these challenges and achieve automation, this paper presents a new localization method that aligns an image to a BIM by detecting and matching perspectives of the image and the BIM. The results demonstrate the potential for enabling automated visual data collection (as-built aligned with as-planned) for performance monitoring.}, booktitle={Computing in Civil Engineering 2017}, author={Boroujeni, K. A. and Han, K.}, year={2017}, month={Jun}, pages={171–178} }