Table 1 Semantic segmentation accuracy.

From: Research on multi-source data fusion and high-precision mapping method for complex landforms based on computer vision

Category

Building

Road

Vegetation

Water body

Ground

Vehicle

mIoU%

Image segmentation

93.1

91.5

88.7

95.2

89.4

82.3

90.0

Point cloud segmentation

94.5

92.8

90.1

-

91.2

85.6

90.8