Single Camera Worker Detection, Tracking and Action Recognition in Construction Site

Activity sampling in the construction industry: a review and research agenda

“…Based on the existing studies, the computer vision-based tracking method for application onsite is considered immature due to several technical issues, such as occlusion, background clutter and camera placement, which are still under extensive research (Han et al, 2013;Ishioka et al, 2020). Existing studies show that computer vision-based methods can only accurately predict labour activities under laboratory or controlled environments, with limited distances from the camera.…”

“…Computer vision predicts the labour activities from the captured video by extracting human body silhouettes (Bai et al, 2012), images contours or shapes, spatiotemporal features (Luo et al, 2018), dense trajectories (Yang et al, 2015(Yang et al, , 2016 and body key joints features (Han et al, 2013;Liu et al, 2017). Since this feature technique interprets fine motions of labour, the labour activities are commonly categorised at level 3 (Ishioka et al, 2020;Yang et al, 2015Yang et al, , 2016. As summarised in Table 3, computer vision-based activity sampling has been conducted onsite for different construction activities such as formwork (Luo et al, 2018), rebar installation (Bai et al, 2012), concreting, carpentry (Liu and Golparvar-Fard, 2015b), scaffolding (Ying et al, 2021), bricklaying (Roberts et al, 2020) and several common construction activities.…”

“…Even in similar workflows, different labours also present different working postures, making it challenging to classify activities (Bai et al, 2012). Besides, body parts occlusion by tools and site environment causes reduced activity recognition accuracy (Han et al, 2013;Ishioka et al, 2020). Hence, extensive efforts are still required to resolve technical issues such as joint body occlusion (Liu et al, 2017), cluttered construction sites (Yang et al, 2016), computational time and cost and varying postures that cause reduced detection accuracy.…”

Development of Action-Recognition Technology Using LSTM Based on Skeleton Data

“…However, depending on the construction site, it may become difficult to setting multiple devices in a limited space. On the other hand, image-based action recognition approaches have got significant results that can recognize the basic work actions of construction workers in workplace (5,6) . Although there is a problem that training data is insufficient due to the worker's body is hidden and action features cannot be acquired correctly when detect object carrying large goods and interacting Fig.…”

Идентификация Поведенческих Паттернов

“…Поскольку не все эти преступления напрямую направлены против людей, в[8] на его основе был собран набор данных, содержащий подмножество преступлений, связанных с людьми.Также алгоритмы видеоаналитики могут быть использованы на промышленных предприятиях. В[9][10][11][12] представлены варианты практического применения распознавания действий на строительной площадке. Поскольку представленные наборы данных содержат типовые действия, происходящие на строительной площадке, анализ техники безопасности с применением этих наборов затруднителен, поскольку отсутствуют классы аномальных действий.Кроме того, анализ видеоданных может быть полезен в электроэнергетике.…”