An Ontology-Based Framework for Complex Urban Object Recognition through Integrating Visual Features and Interpretable Semantics

Table 1

Low-level features and computing methods.

Features	Parameters (rule)	Parameters (test)
Color features (each waveband of optical satellite image and)	Max	Mean
	Min	Standard deviation
	Standard deviation (value range)	Skewness
	Skewness (value range)	Kurtosis
	Kurtosis (value range)
Height features (LiDAR height raster)	Max	Mean
	Min	Standard deviation
	Standard deviation (value range)	Skewness
	Skewness (value range)	Kurtosis
	Kurtosis (value range)
Textural features (fused waveband of optical satellite image and)	Normalized symmetrical matrix of grey-level co-occurrence matrix (GLCM)	Normalized symmetrical matrix of GLCM
	GLCM features	GLCM features
Height textural features (LiDAR height raster)	Normalized symmetrical matrix of grey-level co-occurrence matrix (GLCM)	Normalized symmetrical matrix of GLCM
	GLCM features	GLCM features
Shape features (fused waveband of optical satellite image)	Segmentation fusion [51]	Segmentation fusion [51]
Height shape features (LiDAR height raster)	Region proposal generated by aspect difference	Region proposal generated by aspect difference

There are four normalized symmetrical matrixes of GLCM in north, east, south, and west directions, respectively. GLCM features are computed based on each normalized symmetrical matrix.