Detection Method for Ancient City Wall Diseases Based on Spectral Information Expansion

Yuyang Xing; Huiqin Wang; Ke Wang; Zhan Wang; Yuan Li

doi:10.3788/LOP241975

Laser & Optoelectronics Progress, Volume. 62, Issue 10, 1030001(2025)

Detection Method for Ancient City Wall Diseases Based on Spectral Information Expansion

Yuyang Xing¹, Huiqin Wang^1、*, Ke Wang¹, Zhan Wang², and Yuan Li³

¹College of Information and Control Engineering, Xi'an University of Architecture and Technology, Xi'an 710055, Shaanxi , China

²Shaanxi Institute for the Preservation of Cultural Heritage, Xi'an 710075, Shaanxi , China

³Xi'an Museum, Xi'an 710074, Shaanxi , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(16)

Fig. 1. Structure diagram of 2D spatial feature extraction module

Download full size

Fig. 2. Structure diagram of 3D spatial-spectral feature expansion module

Download full size

Fig. 3. Structure diagram of sparse spectral channels city wall disease detection algorithm

Download full size

Fig. 4. 16 bands multispectral image data of city walls

Download full size

Fig. 5. Spectral images of city walls. (a) Multispectral image data of city walls in 5 bands; (b) color image of city walls

Download full size

Fig. 6. Ablation experiment visualization. (a) Evaluation metric radar chart; (b) complexity analysis visualization

Download full size

Fig. 7. Multispectral images obtained through spectral data expansion. (a) Reference standard image; (b) Conv+ $L_{M S E}$ ; (c) RCAM+ $L_{M S E}$ ; (d) Conv+ ${L_{M S E +}}_{S A M}$ ; (e) RCAM+ ${L_{M S E +}}_{S A M}$

Download full size

Fig. 8. Visualization of classification accuracy

Download full size

Fig. 9. Reflectance curves of various damage types

Download full size

Fig. 10. Confusion matrix for disease detection in proposed model

Download full size

Fig. 11. Visualization of city wall disease detection results under different spectral information expansion algorithms. (a) Color image of target area; (b) CNN; (c) LRDAN+CNN; (d) CTJN+CNN; (e) DCN+CNN; (f) proposed model

Download full size

Table 1. Comparison of ablation experiment evaluation results
View table
Table 1. Comparison of ablation experiment evaluation results
Index Module
Conv+ $L_{M S E}$ Conv+ ${L_{M S E +}}_{S A M}$ RCAM+ $L_{M S E}$ RCAM+ ${L_{M S E +}}_{S A M}$
RMSE 0.1853 0.1531 0.1735 0.1141
MSSIM 0.6207 0.6320 0.5696 0.7869
SAM 0.1995 0.1592 0.1354 0.0646
UIQI 0.7180 0.8469 0.7854 0.9370

Table 2. Comparison of spectral information expansion quality among different algorithms
View table
Table 2. Comparison of spectral information expansion quality among different algorithms
Metric LRDAN CTJN DCN Proposed Best
RMSE 0.3049 0.2639 0.1439 0.1141 0
MSSIM 0.5395 0.5008 0.6424 0.7869 1
SAM 0.4792 0.4444 0.1687 0.0646 0
UIQI 0.6636 0.6570 0.8498 0.9370 1

Table 3. Visualization of spectral information expansion results from different algorithms
View table
Table 3. Visualization of spectral information expansion results from different algorithms
Reference image LRDAN CTJN DCN Proposed

Table 4. Comparison of classification accuracy for surface disease detection on city walls under different spectral information expansion algorithms

View table

Table 4. Comparison of classification accuracy for surface disease detection on city walls under different spectral information expansion algorithms

Category	CNN	LRDAN+CNN	CTJN+CNN	DCN+CNN	Proposed
Kappa	0.7285	0.7759	0.8006	0.8301	0.9212
Biology	82.13%	85.78%	87.69%	97.05%	97.18%
Salting	69.04%	77.69%	85.84%	89.60%	92.77%
Spalling	64.86%	84.85%	85.84%	86.31%	88.28%
Basal brick	84.03%	87.36%	92.61%	93.30%	94.84%
OA	80.11%	92.69%	86.71%	88.85%	94.88%

Table 5. Statistical results of the proportions of disease areas

View table

Table 5. Statistical results of the proportions of disease areas

	CNN	LRDAN+CNN	CTJN+CNN	DCN+CNN	Proposed
Biology	0.63 m²/22.03%	0.77 m²/26.92%	0.71 m²/24.82%	0.81 m²/28.32%	0.66 m²/23.08%
Salting	0.22 m²/7.69%	0.32 m²/11.19%	0.25 m²/8.74%	0.30 m²/10.49%	0.36 m²/12.59%
Spalling	1.19 m²/41.61%	0.48 m²/16.78%	0.33 m²/11.54%	0.38 m²/13.29%	0.12 m²/4.20%

Tools

Get Citation

Copy Citation Text

Yuyang Xing, Huiqin Wang, Ke Wang, Zhan Wang, Yuan Li. Detection Method for Ancient City Wall Diseases Based on Spectral Information Expansion[J]. Laser & Optoelectronics Progress, 2025, 62(10): 1030001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Spectroscopy

Received: Sep. 10, 2024

Accepted: Nov. 5, 2024

Published Online: May. 12, 2025

The Author Email: Huiqin Wang (hqwang@xauat.edu.cn)

DOI:10.3788/LOP241975

CSTR:32186.14.LOP241975

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Comparison of ablation experiment evaluation results

Table 1. Comparison of ablation experiment evaluation results

Table 2. Comparison of spectral information expansion quality among different algorithms

Table 2. Comparison of spectral information expansion quality among different algorithms

Table 3. Visualization of spectral information expansion results from different algorithms

Table 3. Visualization of spectral information expansion results from different algorithms

Table 4. Comparison of classification accuracy for surface disease detection on city walls under different spectral information expansion algorithms

Table 4. Comparison of classification accuracy for surface disease detection on city walls under different spectral information expansion algorithms

Table 5. Statistical results of the proportions of disease areas

Table 5. Statistical results of the proportions of disease areas