Fine Classification of Tree Species Based on Improved U-Net Network

Fig. 4. LiDAR point cloud data and true-color images for study area 1 and 2. (a) LiDAR point cloud data for study area 1; (b) LiDAR point cloud data for study area 2; (c) true-color image for study area 1; (d) true-color image for study area 2

Download full size

Fig. 5. Comparison of classification results of CA-U-Net and U-Net

Download full size

Fig. 6. Model training process. (a) U-Net; (b) CA-U-Net

Download full size

Fig. 7. Proportion of different samples

Download full size

Fig. 8. Comparison of classification results of CB-CA-U-Net and CA-U-Net

Download full size

Table 1. Comparison of classification accuracy between different networks

View table

Table 1. Comparison of classification accuracy between different networks

Classification	FCN		SegNet		U-Net		CA-U-Net
Classification	R /%	F₁ /%	R /%	F₁ /%	R /%	F₁ /%	R /%	F₁ /%
Silver birch	97.26	82.51	95.95	86.35	98.02	97.43	95.10	96.52
Camphor	49.66	63.67	62.68	57.03	99.92	98.87	99.92	99.20
Beech	97.01	94.21	97.50	96.39	96.62	96.16	98.04	96.88
Hackberry	64.25	75.89	13.71	22.11	88.76	92.94	91.44	93.82
Banyan tree	93.32	94.60	96.46	91.39	98.35	91.24	98.09	95.66
Mahogany	96.65	95.41	92.04	95.39	79.03	87.56	95.43	97.32
Small-leaved olive	97.86	88.29	97.56	96.48	98.18	97.92	98.20	98.95
OA /%	88.24		84.81		93.49		96.80
Kappa	0.86		0.82		0.92		0.96

Table 2. Comparison of performance of CA-U-Net and CB-CA-U-Net

View table

Table 2. Comparison of performance of CA-U-Net and CB-CA-U-Net

Classification	CA-U-Net		CB-CA-U-Net
Classification	R /%	F₁ /%	R /%	F₁ /%
Camellia oleifera	91.62	77.75	81.37	82.39
Chinese fir	89.33	87.24	85.22	90.57
Small-leaved olive	67.07	77.24	68.20	73.04
Pinus massoniana	92.07	86.91	90.59	91.25
Bamboo	54.13	57.37	71.76	67.94
Camphor	83.76	85.11	91.23	84.05
Beech	65.15	65.52	80.87	82.39
OA /%	83.93		87.96
Kappa	0.80		0.85

Tools

Get Citation

Copy Citation Text

Yulin Cai, Hongzhen Gao, Xiaole Fan, Huiyu Xu, Zhengjun Liu, Geng Zhang. Fine Classification of Tree Species Based on Improved U-Net Network[J]. Laser & Optoelectronics Progress, 2025, 62(2): 0228002

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Remote Sensing and Sensors

Received: Apr. 26, 2024

Accepted: May. 24, 2024

Published Online: Jan. 6, 2025

The Author Email:

DOI:10.3788/LOP241175

CSTR:32186.14.LOP241175

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology