This paper proposes a new three-dimensional (3D) circle detection algorithm by employing an available two-dimension (2D) circle detector to locate and measure circles in 3D point cloud data (PCD). For 3D PCD collection, a laser-vision-based scanning system was developed and implemented based on the principle of laser-vision triangulation. A direct calibration method was also introduced to calculate the relationship between the image coordinates and the real coordinates in PCD acquisition process. From the obtained 3D PCD, the normal vectors of the scanned surface are calculated. Then, a 2D normal surface image is defined by projecting the 3D PCD along a selected normal vector. The center and radius of 2D circles on the normal surface image are located and measured by applying the 2D circle detection. Next, these center coordinates of detected 2D circles are reverse transformed into 3D PCD coordinates. Finally, the center coordinates, radius, and orientation of 3D circles in the point cloud data are determined. The proposed 3D circle detection method was tested by performing the experiment for locating the center and measuring the radius of circular holes on a sample object. The experimental results showed that the proposed 3D circle detection method and the designed laser-vision-based scanning system can locate and measure circles in the 3D PCD of the scanned object with high accuracy and reliability.