Building outer wall defect detection method, device and equipment based on semantic mask guidance

Abstract

This invention relates to the field of wall defect recognition technology, and discloses a method, apparatus, and equipment for detecting building exterior wall defects based on semantic mask guidance. The method includes: performing semantic segmentation processing on a target visible light image to obtain first location information of a wall region or second location information of a non-wall region in the target visible light image; processing the target visible light image based on the first or second location information to obtain a target wall mask; determining a first target region image corresponding to the wall region in the target visible light image based on the target wall mask; determining a second target region image corresponding to the wall region in a target thermal infrared image based on the target wall mask; performing defect detection on the building exterior wall based on the first target region image to obtain first defect information; and performing defect detection on the second target region image to obtain second defect information, effectively improving the accuracy and reliability of building exterior wall defect recognition.

Images