A study of visual attention and search features for construction site hazard identification incorporating VR-eye movement
To investigate the visual attention and search characteristics involved in recognizing hidden hazards at construction sites,and to enhance the level of recognition performance,we conducted an experiment in which Virtual Reality(VR)and eye-tracking technology were integrated to simulate construction scenes.Eye-movement data were collected from two groups of participants with different levels of experience in recognizing hidden hazards.Statistical methods were employed to screen and collate eye-movement indicators,such as the first gaze time,and compare the recognition performance of the two groups.Differences in the cumulative gaze time,gaze percentage,and other eye-movement indicators were analyzed to reveal the intrinsic connection between the level of experience in recognizing hidden hazards and the allocation of attention resources.Using the eye-movement trajectory matching method,the first visual search trajectories of the two groups were clustered,and the differences in search habits under the influence of the experience of recognizing hidden hazards were investigated by combining the spatial layout of the construction site.The results of the experiment showed that the level of empirical knowledge played a key role in driving the performance of hazard recognition,but the effect of empirical knowledge was weakened by a low level of hazard expertise.The attentional resources of the subjects were mostly focused on irrelevant elements with visual saliency characteristics,and only a small proportion of attentional resources were allocated to hazards in the virtual construction scenario.However,subjects with a higher level of empirical knowledge were more sensitive to hidden danger information and allocated a higher proportion of attention resources to hazards at the construction site.For the visual search habit of hidden hazards in the virtual scene,the expert group mostly used a detailed search of one area before switching to the next,while the novice group repeatedly searched the same area.
safe social engineeringhazard identificationVirtual Reality(VR)eye-movement testidentifying features
NETL
NSTL
万方数据
维普
