The invention discloses a building detection method and device based on 
point cloud data analysis, and the device is used for achieving the method, and the method comprises the steps: building a building reference index 
library which at least comprises a wall surface, a ground, a ceiling horizontal 
reference surface and the 
angular deviation and spacing deviation of the wall surface, the ground and the ceiling horizontal 
reference surface; acquiring indoor actual 
measurement point cloud data of the target building acquired by 
laser scanning; converting a coordinate 
system of the actual 
measurement point cloud data into a local coordinate 
system; identifying surface data in the 
point cloud data, wherein the surface data at least comprises wall surface data, ground data and ceiling data; fitting the data of each surface by adopting a least square method to generate a corresponding fitting surface; calculating the angle difference between the fitting surface and the 
reference surface, andif the angle difference exceeds the angle deviation, marking the surface as an unqualified surface; and calculating the distance between each actual 
measurement point in the surface data and the fitting surface of the actual measurement point, and if the distance exceeds the distance deviation range, marking the actual measurement point as an unqualified point, thereby generating an actual measurement result. The detection method does not need manual intervention, and is efficient, rapid and accurate in data.