Research on Fully Automatic Template Calibration Method Based on Machine Vision Guidance.

In order to address the significant manual costs incurred during template measurements due to prism placement and manual observations, as well as to mitigate the safety hazards associated with working aloft, this paper proposes a vision-based AI-driven continuous template automatic measurement approach. Employing an image total station as the primary sensing device, the method utilizes machine vision algorithms to replace manual observation methods for the automated detection and measurement of continuously rising templates at dam construction sites. This approach enhances the efficiency of dam construction layout and reduces the workload for measurement personnel. The method incorporates concentric circles with numerical markings as collaborative targets to replace prisms, and employs a circle center detection algorithm based on cross ratio invariance for precise identification and positioning of measurement points. Additionally, a DP controller is applied to achieve accurate alignment of the image total station. Experimental measurements were conducted at the TB dam pouring site using the image total station Leica TM50I for automated measurements. The results demonstrate that within a testing distance of 70 to 131 meters, the error between this method and traditional manual measurements is within 5 mm. [ABSTRACT FROM AUTHOR]