In this study, we present an innovative computer-vision-based displacement monitoring (CVDM) instrument, which includes a microcomputer, camera module, telescopic lens, and chessboard. The computer vision technology is employed to detect ground movements with an ultralong-distance, long-term, solar-powered, and real-time monitoring system. The average standard deviations and resolutions of three-axis displacements in the CVDM are 0.06 pixels and 0.01 cm at the 30 cm distance and 0.95 pixels and 0.10 cm at the 50 m distance, respectively, in experimental tests. The CVDM system can reach an ultralong distance of 1000 m in the field. The standard deviation and resolution of the ultralong-distance test are 0.33 pixels and 0.27 cm, respectively. Then, the CVDM system was installed inside a self-designed enclosure for long-term and real-time monitoring on a slope of the Jiufenershan landslide area with a 33 m distance between the enclosure and a chessboard. The CVDM system in the area can detect four frames of image recognition per second and transmit three-axis relative displacement and image data every 10 min with a 5G network for 24-h monitoring. Finally, the CVDM system is the artificial intelligence of things (AIoT) solution for computer-vision-based, economically efficient, and energy-saving slope monitoring.