AbstractWhen an electromagnetic wave passes through the atmosphere, due to variation of atmospheric parameters, its propagation path is usually altered and the propagation speed also changes. Therefore, errors will be induced in triangulation measurement with electronic surveying robots, delimiting the accuracy of trigonometric leveling. Targeting at improving the accuracy of trigonometric leveling, a correction model based on atmospheric refraction coefficients is proposed in this article. First, based on the meteorological parameters collected on-site and Rüeger equation, influence of the meteorological parameters (e.g., temperature, atmospheric pressure, humidity, and CO2) on the refractive index are analyzed. Furthermore, gradient models of temperature, humidity, and atmospheric pressure with reference to height are established. In order to build the atmospheric gradient models of the optical path, height difference between the optical path and the ground is estimated with a precise digital elevation model generated by laser scanning. On this basis, the vertical atmospheric refraction coefficient model is proposed and used to correct the trigonometric leveling data measured by surveying robots. By conducting experimental studies over two open-pit iron mines with large topographic relief and comparing with first-class leveling data, reliability of the proposed atmospheric refraction coefficient model is assessed. The outcome of this article has demonstrated that by adopting the proposed model, accuracy of trigonometric leveling can meet the requirements of second-class leveling.