Semantic Segmentation Method for Remote Sensing Images Based on Improved DeepLabV3+

Fig. 7. Comparison of experimental results. (a) Original images; (b) image labels; (c) FCN processing results; (d) U-Net processing results; (e) SegNet processing results; (f) DeepLabV3+ processing results; (g) AFSM-Net processing results

Download full size

Table 1. Comparison of evaluation results of ablation experiment on CFF test set
View table
Table 1. Comparison of evaluation results of ablation experiment on CFF test set
ID Baseline CA SA MIoU /%
a √ × × 82.58
b √ √ × 84.26
c √ × √ 85.02
d √ √ √ 86.78

Table 2. Experimental results of different cutting ratios
View table
Table 2. Experimental results of different cutting ratios
ID k MIoU /%
a 1 79.65
b 2 86.40
c 3 83.64
d 4 80.02

Table 3. Classification accuracy of different algorithms

View table

Table 3. Classification accuracy of different algorithms

Algorithm	Vegetation	Building	Water	Road	OA	MIoU
FCN	70.22	71.55	69.25	70.22	70.31	70.84
U-Net	81.85	80.26	80.32	79.84	80.57	75.26
SegNet	83.46	83.28	82.84	82.78	83.09	79.45
DeepLabV3+	88.26	88.53	87.00	88.45	88.56	82.48
AFSM-Net	90.46	90.56	89.84	89.64	90.13	86.42

Tools

Get Citation

Copy Citation Text

Zhipeng Su, Jingwen Li, Jianwu Jiang, Yanling Lu, Ming Zhu. Semantic Segmentation Method for Remote Sensing Images Based on Improved DeepLabV3+[J]. Laser & Optoelectronics Progress, 2023, 60(6): 0628003

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites