Modeling object identification and tracking errors on automated spatial safety assessment of earthmoving operations
Recent research studies have been conducted for automating the safety assessment process in order to identify risks and safety hazards on a job site without human intervention. Regardless of the benefits of automated assessment, safety planners still face challenges selecting applicable devices, methods, and algorithms for safety assessment. This is due to the fact that (1) such devices, methods, and algorithms typically have measurement and processing errors, (2) construction operations and sites are unique and complex, and (3) the impact of the errors is different depending on workspaces. The primary objective of this research is to develop an error impact analysis method to model data collection and data processing errors caused by image-based devices and algorithms and to analyze the impact of the errors for spatial safety assessment of earthmoving and surface mining activities. The literature review revealed the possible causes of accidents on earthmoving activities, investigated the spatial risk factors of these types of accident, and identified spatial data needs for safety assessment based on current safety regulations. Image-based data collection devices and algorithms for safety assessment were then evaluated. Analysis methods and rules for monitoring safety violations were also discussed. A testbed to model and simulate workspaces and related spatial safety violations was finally designed. Using the testbed, the impacts of image-based algorithm and device errors―more specifically, object identification and tracking errors―on the data collected and processed were investigated for the safety planning purpose. Field experiments assessed the feasibility of automated spatial data collection and analysis methods. Industrial project and safety experts verified the proposed safety rules and the testbed design. Computer simulations were conducted for testing the proposed testbed. The testbed was used to model several earthmoving operation scenarios, detect simulated safety violations using safety rules, and finally evaluate the impact of different object identification and tracking errors on the safety analyses. The result of this research could be used for improving site safety assessment and planning by assisting safety planners to understand workspaces and to evaluate errors related to the use of different image-based technologies for safety assessment of earthmoving and surface mining activities.