Roughness is a characteristic of a surface that is a function of its geometry. Ultrasonic sensing in air provides range, area, and angle information because the surface geometry determines the characteristics of the echo. The authors introduce the “spatial-angle-filter model” to explain the impact of surface roughness on the echo. On the basis of this model, they design a set of features for use in classifying surfaces. The quality of the features and the classification is measured with the Mahalanobis distance. The resultant system is able to achieve 99.73% classification of a set of 12 surfaces using five features.