基于级联密集残差网络的温度场高分辨率重建

doi:10.11996/JG.j.2095-302X.2023020216

摘要/Abstract

摘要：

温度分布的高质量测量对工业生产具有重要意义。作为一种非侵入性的测量方法，声学层析成像(AT)被认为是一种有前景的温度分布可视化技术。为改善重建质量，提出了一种虚拟观测(VO)结合级联密集残差网络(CDRNN)的二阶段温度场高分辨率重建算法。首先通过 VO 算法获取粗网格温度场，然后搭建 CDRNN 预测细网格温度信息。采用 VO 算法获取 AT 反问题的总体最小二乘解，缓解了声波路径弯曲引起的重建误差。引入双输入补偿策略增加了 CDRNN 对原始测量信息的利用率，提高了网络稳定性，通过设置子网络简化网络结构，并采用密集连接与残差连接改善网络信息流，同时引入亚像素卷积层降低网络计算维度，提高重建精度。对典型温度场模型进行数值模拟，并与 Landweber 迭代法、ART 算法、ART-NN 算法和 VO 算法进行比较。结果表明 VO-CDRNN 算法的平均相对误差和均方根误差分别为 0.44%和 0.68%，重建结果均优于其他算法。

关键词:

"> 声学层析成像, 高分辨率重建, 虚拟观测, 残差连接, 密集连接, 亚像素卷积层 ,

Abstract:

The high-quality measurement of temperature distribution is of great importance for industrial production. As a non-invasive measurement method, acoustic tomography (AT) is considered as a promising technique for the visualization of temperature distribution. To enhance the reconstruction quality, a two-stage high-resolution reconstruction algorithm was proposed for temperature field based on virtual observation (VO) and cascaded dense residual network (CDRNN). Firstly, the temperature field of coarse grid was obtained by the virtual observation algorithm, and then the CDRNN was built to predict the fine grid temperature distribution. The VO algorithm was used to achieve the overall least squares solution of the AT inverse problem, thereby reducing the reconstruction error caused by the bending of the acoustic path. Additionally, a dual-input compensation strategy was introduced to increase the utilization of the original measurement information by the CDRNN, and to improve the network stability. The network structure was streamlined by setting up sub-networks, dense connections and residual connections were also employed to improve network information flow. Sub-pixel convolutional layers were introduced to decrease network computing dimensions and boost reconstruction accuracy. Finally, the effectiveness of the algorithm was verified using a variety of simulated temperature field models. Through the numerical simulation of a typical temperature field model and comparison with the Landweber iterative method, ART algorithm, ART-NN algorithm, and VO algorithm, it was found that the average relative error and root mean square error of the VO-CDRNN algorithm were 0.44% and 0.68%, respectively, thus achieving better reconstruction results than other algorithms.

Key words:

"> acoustic tomography, high-resolution reconstruction, virtual observation, residual connection, dense connection, sub-pixel convolutional layer

中图分类号:

TP 391.4

张立峰, 李晶. 基于级联密集残差网络的温度场高分辨率重建[J]. 图学学报, 2023, 44(2): 216-224.

ZHANG Li-feng, LI Jing. High resolution reconstruction of temperature field based on cascaded dense residual network[J]. Journal of Graphics, 2023, 44(2): 216-224.

参考文献

[1] 李雄兵, 胡宏伟, 倪培君, 等. 基于图像处理的超声探头声场测量[J]. 工程图学学报, 2011, 32(4): 27-30. LI X B, HU H W, NI P J, et al. Acoustic field measurement of ultrasonic probe based on image processing[J]. Journal of Engineering Graphics, 2011, 32(4): 27-30 (in Chinese).

[2] PAL S, LIN F S, HSIEH C C, et al. An acoustic hotspot tracking algorithm for highly centralized gas temperature distribution[J]. IEEE Trans Ultrason Ferroelectr Freq Control, 2021, 68(4): 1370-1379.

[3] 刘厦, 刘石. 基于声学层析成像的炉内温度场重建算法研究[J]. 动力工程学报, 2017, 37(7): 525-532, 568. LIU S, LIU S. Research on temperature distribution reconstruction of a boiler furnace based on acoustic tomography[J]. Journal of Chinese Society of Power Engineering, 2017, 37(7): 525-532, 568 (in Chinese).

[4] RADAJEWSKI M, DECKER S, KRÜGER L. Direct temperature measurement via thermocouples within an SPS/FAST graphite tool[J]. Measurement, 2019, 147: 106863.

[5] YAO K, LIU Y C, ZHOU Y P, et al. Influence of thermal contact resistance on thermal conductivity measurement with a high-temperature guarded hot plate apparatus[J]. International Journal of Thermophysics, 2020, 41(2): 1-12.

[6] XU L, XUE R, LI Y, et al. FPGA-based real-time implementation of temperature measurement via tunable diode laser absorption spectroscopy[J]. IEEE Sensors Journal, 2018, 18(7): 2751-2758.

[7] 吕晓光, 王明泉, 李光亚. 合成孔径聚焦超声成像在混凝土探伤中的应用研究[J]. 图学学报, 2014, 35(6): 946-949. LV X G, WANG M Q, LI G Y. Application of synthetic aperture focusing technique in ultrasonic detection of concrete[J]. Journal of Graphics, 2014, 35(6): 946-949 (in Chinese).

[8] YU X, LIN S, HONG X, et al. Reconstructing underwater temperature spatial distribution using ultrasound tomography[J]. IEEE Access, 2021, 9: 103670-103676.

[9] LIU S, LIU S, REN T. Acoustic tomography reconstruction method for the temperature distribution measurement[J]. IEEE Transactions on Instrumentation and Measurement, 2017, 66(8): 1936-1945.

[10] YANG W Q, SPINK D M, YORK T A, et al. An image-reconstruction algorithm based on Landweberʹs iteration method for electrical-capacitance tomography[J]. Measurement Science and Technology, 1999, 10(11): 1065-1069.

[11] ZHANG S, SHEN G, AN L, et al. Online monitoring of the two-dimensional temperature field in a boiler furnace based on acoustic computed tomography[J]. Applied Thermal Engineering, 2015, 75: 958-966.

[12] LIU S, LIU S, REN T. Ultrasonic tomography based temperature distribution measurement method[J]. Measurement, 2016, 94: 671-679.

[13] ZHAO G, LEI B, YE G, et al. Improved eccentricity self-detection method based on least square algorithm for polar coordinate encoder[J]. IEEE Sensors Journal, 2021, 21(23): 26902-26911.

[14] DOKHANCHI N, ARNOLD J, VOGEL A, et al. Measurement of indoor air temperature distribution using acoustic travel-time tomography: optimization of transducers location and sound-ray coverage of the room[J]. Measurement, 2020, 164: 107934.

[15] LIU S, LIU S, TONG G. Reconstruction method for inversion problems in an acoustic tomography based temperature distribution measurement[J]. Measurement Science and Technology, 2017, 8(11): 115005.

[16] 刘厦, 刘石, 任婷. 基于 SA-ELM 的声学层析成像温度分布重建算法[J]. 化工学报, 2017, 68(6): 2434-2446. LIU S, LIU S, REN T. SA-ELM based method for reconstructing temperature distribution in acoustic tomography measurement[J]. CIESC Journal, 2017, 68(6): 2434-2446 (in Chinese).

[17] SVENŠEK D. Ray-tracing semiclassical (RTS) low frequency acoustic modeling validated for local and extended reaction boundaries[J]. Journal of Sound and Vibration, 2018, 437: 1-15.

[18] 王乐洋, 于冬冬. 病态总体最小二乘问题的虚拟观测解法[J]. 测绘学报, 2014, 43(6): 575-581. WANG L Y, YU D D. Virtual observation method to ill-posed total least squares problem[J]. Acta Geodaetica et Cartographica Sinica, 2014, 43(6): 575-581 (in Chinese).