Quantitative Evaluation of CHT and GHT for Column Detection under Different Conditions of Data Quality

Quality control and project monitoring are topics of interest in the field of architectural-engineering-construction/facility management. The need to automatise the process and analyze data acquired with terrestrial laser scanners leads to optimum ways to perform the data acquisition guaranteeing the quality standards. This paper is focussed on the quantitative evaluation of a column detector based on the generalized hough transform (GHT) and the circle hough transform (CHT). Several quality conditions are considered, with the main goal to be able to determine the minimum requirements needed to be established during the data surveying. Elements such as completeness, point density, noise, and accuracy are simulated in laboratory conditions to perform the evaluation. Size and center position of round and rectangular cross-section columns are analyzed, reaching to a final conclusion that columns must be acquired with at least 30–40% of completeness. Noise effect is centered in the position of the column applying an error between 4 and 8 pixels. Point cloud resolution does not affect the results, whereas the projected image is stable and well distributed. The evaluation also showed a better performance of the Hough Transform on round than rectangular cross-section columns, primarily because the last ones are more sensitive to the amount of sides acquired and completeness values.