1 School of Traffic & Transportation, Chongqing Jiaotong University, Chongqing 400074, China
2 School of Mechatronics and Vehicle Engineering, Chongqing Jiaotong University, Chongqing 400074, China
Show less
文章历史+
Received
Published
Revised
2025-05-27
2026-06-08
2025-11-26
Online Date
2026-08-07
PDF (16550K)
摘要
针对公路隧道口光照突变导致的自动驾驶车辆视觉适应性问题,提出一种非下采样剪切波变化(nonsubsampled shearlet transform,NSST)域下渐进式的红外与可见光图像融合方法。该方法构建多尺度、多方向的层级分解结构,获得从粗到细的低频结构信息与多方向高频细节信息;对低频系数,创新结合分块处理机制与ADMM(alternating direction method of multipliers)优化算法改进卷积稀疏表示,有效解决全局特征一致性问题;对高频系数,引入光谱边缘信息并基于新活性度量方法实现对融合图像的纹理细节优化;最后,利用非下采样轮廓波逆变换重构融合图像。实验选取G22宁夏六盘山隧道进行融合实验验证,结果表明:所提方法与传统算法相比,互信息(mutual information, MI)、边缘信息传递量(transfer of edge information, QAB/F)和自然图像质量评价指标(natural image quality evaluator,NIQE)指标值至多提高0.98、0.22和2.2,运行时间至多提升17.5%,克服了传统算法的“块效应”和“伪吉布斯”效应,增强图像间相关性。将融合图像输入YOLO11s目标检测网络,mAP@0.5达到92.45%,验证了该方法在下游感知任务中的实际增益。综上,所提方法有助于提升自动驾驶系统对低照度、高动态范围环境的感知精度,同步改善行车安全性与通行效率。
Abstract
Aiming at the visual adaptation challenges of autonomous vehicles caused by abrupt illumination changes at highway tunnel portals, a progressive infrared and visible image fusion method in the nonsubsampled shearlet transform (NSST) domain was proposed. The method constructed a multi-scale, multi-directional hierarchical decomposition to obtain coarse-to-fine low-frequency structural information and multi-directional high-frequency details. For low-frequency coefficients, convolutional sparse representation was enhanced by integrating a block-wise processing mechanism with alternating direction method of multipliers (ADMM) optimization, effectively improving global feature consistency. For high-frequency coefficients, spectral edge information was incorporated and text ure details were preserved through a novel activity measure. The fused image was reconstructed via inverse non-subsampled contourlet transform. Experiments on real data from the G22 Ningxia Liupan Mountain Tunnel show that, compared with traditional algorithms, the proposed method improves mutual information (MI), transfer of edge information (QAB/F), and natural image quality evaluator (NIQE) by up to 0.98, 0.22, and 2.2, respectively, while reducing runtime by up to 17.5%. It effectively mitigates the “block effect” and “pseudo-Gibbs effect” and enhances inter-image correlation. When fed into the YOLO11s object detection network, the fused images achieve a mAP@0.5 of 92.45%, demonstrating clear benefits for downstream perception tasks. In conclusion, the proposed approach is helpful to enhance the perception accuracy of autonomous driving systems in low-illumination, high-dynamic-range environments, thereby improving driving safety and traffic efficiency.
MA Qing-lu, WANG Chen, MA Lian, ZHAO Shu-en.
Illumination Adaptive Progressive Multi-scale Image Fusion for Tunnel Portal[J].
Science Technology and Engineering, 2026, 26(16): 6954-6965 DOI:10.12404/j.issn.1671-1815.2503922
为优化全局问题,将边缘特征扩展到整个区域,提出一种改进算法,将全局优化问题拆分成多个局部问题,通过局部子问题的求解逐步构建整体图像的最终特征表达,实现对图像特征从边缘到整体的有效转移和全局最优化,并利用ADMM(alternating direction method of multipliers)解凸优化,模型变量定义为{ym},如式(7)所示。
LiuQian, WangXuesong, WangChangjun. Readiness research on right turns at intersections based on autonomous driving safe sight zones[J]. China Journal of Highway and Transport, 2025, 38(3): 31-47.
LiuWanjun, ChengYuqian, QuHaicheng. Image self-enhancement de-hazing algorithm combined with generative adversarial network[J]. Journal of System Simulation, 2024, 36(5): 1093-1106.
[10]
XianR, XiongX, PengH, et al. Feature fusion method based on spiking neural convolutional network for edge detection[J]. Pattern Recognition, 2024, 147: 110112.
[11]
ZhaoY, ZhengQ, ZhuP, et al. TUFusion: a transformer-based universal fusion algorithm for multimodal images[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2023, 34(3): 1712-1725.
[12]
YangG, LiJ, LeiH, et al. A multi-scale information integration framework for infrared and visible image fusion[J]. Neurocomputing, 2024, 600: 128116.
[13]
LuoD, LiuG, BavirisettiD P, et al. Infrared and visible image fusion based on VPDE model and VGG network[J]. Applied Intelligence, 2023, 53(21): 24739-24764.
[14]
TangL, XiangX, ZhangH, et al. DIVFusion: darkness-free infrared and visible image fusion[J]. Information Fusion, 2023, 91: 477-493.
[15]
JiangH, LuoA, FanH, et al. Low-light image enhancement with wavelet-based diffusion models[J]. ACM Transactions on Graphics (TOG), 2023, 42(6): 1-14.
YaoRui, WangKai, GuoHaofan, et al. Infrared and visible image fusion based on cross-modal feature interaction and multi-scale reconstruction[J]. Infrared and Laser Engineering, 2025, 54(8): 20250210.
WangAixia, HuAojie, YanAiyun, et al. An image fusion network based on ambient light awareness and infrared feature layer guidance[J]. Control and Decision, 2025, 40(10): 3177-3189.
CaiShuo, YaoXuanshi, TangYuanzhi, et al. Scene-adaptive knowledge distillation-based fusion of infrared and visible light images[J]. Journal of Electronics & Information Technology, 2025, 47(4): 1150-1160.
JiangMai, ShaGuijun, LiNing, et al. Infrared and visible light images fusion based on tetrolet transform and infrared saliency feature extraction[J]. Science Technology and Engineering, 2022, 22(30): 13398-13405.
HanYang, YangHua. An image fusion algorithm based on nonsubsampled shearlet transform coupled with significant information weighting[J]. Science Technology and Engineering, 2021, 21(17): 7224-7229.
SunLei, TangQian, LiaoYipeng, et al. Multi-scale enhancement and color depth codec correction of flotation foam low illumination images[J]. Optics and Precision Engineering, 2025, 33(10): 1609-1626.
LiYingchao, ZhaoZhehao, WangQi, et al. Polarization spectral image fusion method for hybrid backgrounds of ground objects[J]. Chinese Optics, 2024, 17(5): 1098-1111.
[35]
LüF, LiuX, ChenL, et al. Low-amplitude structure recognition method based on non-subsampled contourlet transform[J]. Petroleum Science, 2024, 21(5): 3062-3078.
[36]
TangH, LiuG, QianY, et al. EgeFusion: towards edge gradient enhancement in infrared and visible image fusion with multi-scale transform[J]. IEEE Transactions on Computational Imaging, 2024, 10: 385-398.
XiongZhifei, ShenJianghai. A multi-focus image fusion algorithm based on multidimensional adaptive filtering and poisson fusion[J]. Science Technology and Engineering, 2023, 23(24): 10427-10436.
CuiXiaorong, ShenTao, HuangJianlu, et al. Infrared mid-wave and long-wave image fusion based on FABEMD and improved local energy window[J]. Spectroscopy and Spectral Analysis, 2021, 41(7): 2043-2049.