聲波測溫（Acoustic Tomography）被認為是一種有潛力的氣體溫度分佈重建技術。透過AT所發展的聲波氣體溫度計是一種通過多條路徑的聲速重建氣體介質中的二維溫度分佈技術。近年來，這項技術逐漸取代了製造業中傳統的溫度測量裝置，原因在於聲波氣體溫度計能夠安裝在諸如鍋爐側壁等位置，並且具有抵抗灰塵干擾的能力。然而，該裝置使用的聲音發射裝置對於室內應用（如智慧建築）來說聲響影響過大。因此，使用更創新和更便捷的方法來監測室內溫度分佈已成為一個廣泛討論的議題。一些研究人員嘗試將聲源替換為超聲波傳感器，但此類超聲波氣體溫度計只能處理小的目標區域(Region of Interest，ROI)並且該追踪動態熱點方面的能力存在不確定性。另外，受到超聲波換能器在發射及接收時的硬體限制，使得獲取良好的接收訊號成為重建溫度分布的重要關鍵之一。因此在本研究中將探討如何實現加速重建溫度分布以及優化超聲波氣體溫度計之訊號發射及接收。
在本篇論文中的第一部分將研究如何提高重建速度。通常，AT的重建算法會利用ROI內多條路徑的聲速，來重建該區域的溫度分布。為了提高重建的準確性，增加迭代次數是不可避免的，但這同時可能會增加耗時。此外，現有的重建算法很少考慮利用實際測量的結果來提高其性能。在本論文中，提出了一種使用卷積神經網絡（Convolutional Neural Network, CNN）的重建方法，訓練一個可以從接收聲速中重建氣體溫度分布的機器學習模型。通過調整模型標籤和損失函數，成功訓練了六個模型來重建氣體的溫度分布。其中兩個模型使用理想的溫度分布進行訓練，三個使用重建的溫度分布進行訓練，另一個使用ROI中特定位置的溫度進行訓練。然後，本論文使用了13個具有不同位置的峰值溫度的溫度分布來測試模型在追踪熱點方面的性能，並且與現有的重建方法進行了比較。此外，還使用了三種不同氣體溫度分布的聲速來評估模型的2D可視化性能。其中一個與訓練數據中的溫度分布相似，另外兩個則完全不同。結果表明，使用理想溫度分布訓練的模型能更準確地追踪熱點，使用重建溫度分布訓練的模型在輸入聲速與訓練數據不同的情況下，與先前的重建算法相比性能相似。然而，在輸入聲速與訓練數據不同的情況下，使用理想氣體溫度分布訓練的模型在2D可視化方面的表現較使用重建溫度分布訓練的模型差。這表明了所提出的方法可以成功地從先前的重建算法中學習溫度分布與聲速之間的關係。此外，所提出模型的重建時間為0.109秒，比先前的迭代重建方法減少了96%。因此，本篇論文所提出的重建方法應該可視為可靠及有效的2D氣體溫度分布重建方法。
在第二部分中，本論文提出了一種可用於室內的二維聲波氣體溫度計，可以即時監測大型ROI的二維氣體溫度分佈。通過使用堆疊設計之換能器和預振動方法增強超聲波信號的發射和接收，並且僅憑硬體即可實現所提出的飛行時間計算(Time of Flight, TOF)算法。提出之系統在實驗中極大地減少了計算單元的工作量，能夠在不到3秒的時間內獲取2D溫度分佈。此外，本論文還開發了一個使用者界面，用於顯示和記錄信息，包括整個ROI的二維氣體溫度分佈、所選橫截面的溫度曲線、特定位置的溫度值和每個聲速路徑的速度。在室內測試中，室內溫度分佈的相對標準偏差小於2%。最後，進行了移動熱源的實驗，用以驗證所提出系統的熱點追踪能力。當熱源每10秒移動5厘米和每30秒移動5厘米時，系統能夠成功追踪加熱器的位置。因此，本論文成功證明了所提出的二維氣體聲波溫度計可以快速準確地測量溫度分佈並追踪大型ROI中的熱點，例如智慧建築內的空間。
本論文包含一個基於深度學習的重建演算法與一個基於現場可程式化邏輯閘陣列(Field-Programmable Gate Array , FPGA)的聲波氣體溫度計電路設計，期望未來進一步優化並且整合設計後，對於監測環境的溫度分佈應用可以做出貢獻。

Acoustic tomography (AT) is considered a promising visualization technique for gas temperature distribution (TD). Acoustic gas thermometry developed through AT is a technique that reconstructs the two-dimensional temperature distribution of a gas medium using the acoustic velocities of multiple propagation paths in a region of interest (ROI). In recent years, this technology has gradually replaced traditional temperature measurement devices in combustion industries because acoustic gas thermometers can be installed on side walls and are resistant to dust interference. However, they are too loud for indoor applications such as intelligent buildings. Thus, monitoring indoor temperature distribution with more innovative and easier approaches has become a widely discussed topic. Some researchers have tried to replace the siren sound source with ultrasonic transducers, but such ultrasound gas thermometers can only handle small ROIs or have unknown dynamic hotspot-tracking abilities. Furthermore, due to the hardware limitations of the ultrasonic transducers during emission and reception, obtaining good received signals becomes one of the key factors in reconstructing temperature distributions. Therefore, this dissertation explores methods to accelerate the process of the temperature distribution reconstruction and optimize the signal emission and reception of the ultrasonic gas thermometer.
In the first part of this dissertation would investigate methods to enhance the reconstruction speed. Generally, a temperature map of a region of interest (ROI) is reconstructed with acoustic velocities of multiple routes inside the ROI using an AT reconstruction algorithm. To improve the accuracy of the reconstruction, increasing the number of iterations is inevitable, which may be time-consuming. Besides, existing reconstruction algorithms rarely consider utilizing practical measurements to improve their performance. In this study, a convolutional neural network is proposed to train a machine learning model that can reconstruct a gas TD from acoustic velocities. By adjusting the label and the loss function, a practical training approach was found. Six models were successfully trained to reconstruct a gas TD. Two were trained with the ideal TD, three with a reconstructed TD, and one with temperatures from specific locations. After that, a TD with peak temperatures located in 13 different positions was applied to test the models’ performance in tracking the hot spot. The results were compared with an existing reconstruction method. Besides, the acoustic velocities from three gas TDs were applied to evaluate the 2D visualization performance. One of them was similar to the training data and the other two were different. The results indicated that the models trained with the ideal gas TD could track the hot spot more closely, and the models trained with reconstructed gas TDs had similar performance as the selected reconstruction algorithm. However, the 2D visualization results using models trained with the ideal gas TD were poor compared with models trained with reconstructed TDs when the input acoustic velocities were different from the training data. This indicated that the proposed method could successfully learn the relationship between TD and acoustic velocities from an ordinary reconstruction algorithm. Furthermore, the execution time of the proposed model was 0.109s, which is 96% less than the selected iterative reconstruction method. Consequently, the proposed neural networks should be considered a reliable and efficient 2D gas TD reconstruction methodology.
In the second part, this dissertation proposes an indoor 2D acoustic gas thermometer that can immediately monitor a large ROI's 2D gas temperature distribution. By enhancing the transmitting and receiving of ultrasonic signals with a stacked transducer and pre-vibration method, the hardware alone can implement the proposed time-of-flight calculation algorithm. The system drastically reduced the effort of the computation units and was able to acquire the 2D temperature distribution in our experiment in less than 3 seconds. Additionally, a graphical user interface was developed to display and record the information, including the 2D gas temperature distribution of the entire ROI, the temperature profile of the selected cross-section, the temperature values of specific positions, and the velocity of each acoustic path. The relative standard deviations of acoustic velocities measured at room temperature and the heated environment were less than 2%. Finally, experiments involving a moving heat source were conducted to validate the hotspot-tracking capability of the proposed system. The system tracked the heater’s location successfully when it moved 5 cm per 10 seconds and 30 seconds. Thus, this dissertation demonstrated successfully that the proposed 2D gas acoustic thermometer can quickly and accurately measure the temperature distribution and follow the hotspot within a large ROI, such as an intelligent building's indoor space.
This dissertation proposes a deep learning-based reconstruction algorithm and an FPGA-based circuit design for the acoustic gas thermometer. It is expected that further optimization and integration of the design will contribute to the monitoring of environmental temperature distributions in the future.

[1]S. Kolouri, M. R. Azimi-Sadjadi, and A. Ziemann, "A statistical-based approach for acoustic tomography of the atmosphere," The Journal of the Acoustical Society of America, vol. 135, no. 1, pp. 104-114, 2014.
[2]Q. Kong, G. Jiang, Y. Liu, and M. Yu, "Numerical and experimental study on temperature field reconstruction based on acoustic tomography," Applied Thermal Engineering, vol. 170, p. 114720, 2020.
[3]S. Liu, S. Liu, and T. Ren, "Acoustic tomography reconstruction method for the temperature distribution measurement," IEEE Transactions on Instrumentation and Measurement, vol. 66, no. 8, pp. 1936-1945, 2017.
[4]S. Zhang, G. Shen, and L. An, "Leakage location on water-cooling wall in power plant boiler based on acoustic array and a spherical interpolation algorithm," Applied Thermal Engineering, vol. 152, pp. 551-558, 2019.
[5]S. Zhang, G. Shen, L. An, and G. Li, "Ash fouling monitoring based on acoustic pyrometry in boiler furnaces," Applied Thermal Engineering, vol. 84, pp. 74-81, 2015.
[6]R. Jia, Q. Xiong, G. Xu, K. Wang, and S. Liang, "A method for two-dimensional temperature field distribution reconstruction," Applied Thermal Engineering, vol. 111, pp. 961-967, 2017.
[7]W.-Y. Tsai, H.-C. Chen, and T.-L. Liao, "High accuracy ultrasonic air temperature measurement using multi-frequency continuous wave," Sensors and Actuators A: Physical, vol. 132, no. 2, pp. 526-532, 2006.
[8]S. Zhang, G. Shen, L. An, and Y. Niu, "Online monitoring of the two-dimensional temperature field in a boiler furnace based on acoustic computed tomography," Applied thermal engineering, vol. 75, pp. 958-966, 2015.
[9]C. Lou and H.-C. Zhou, "Deduction of the two-dimensional distribution of temperature in a cross section of a boiler furnace from images of flame radiation," Combustion and Flame, vol. 143, no. 1-2, pp. 97-105, 2005.
[10]C.-Y. Lu, S.-W. Du, and S.-K. Kuo, "Development of an online blast furnace burden profile measuring system," China Steel Technical Report, vol. 23, pp. 25-30, 2010.
[11]Y. S. Semenov, E. Shumelchik, and V. Horupakha, "Expert Module of the Thermal Probes System for Blast Furnace Charging Control," Steel in Translation, vol. 48, pp. 802-806, 2018.
[12]S. Zhang, G. Shen, and L. An, "Online monitoring of furnace exit gas temperature in power plants," Applied Thermal Engineering, vol. 147, pp. 917-926, 2019.
[13]G. Pilato and G. Vassallo, "TSVD as a statistical estimator in the latent semantic analysis paradigm," IEEE Transactions on Emerging Topics in Computing, vol. 3, no. 2, pp. 185-192, 2014.
[14]R. A. Willoughby, "Solutions of ill-posed problems (AN Tikhonov and VY Arsenin)," Siam Review, vol. 21, no. 2, p. 266, 1979.
[15]J. Zhang et al., "Acoustic tomography of two dimensional velocity field by using meshless radial basis function and modified Tikhonov regularization method," Measurement, vol. 175, p. 109107, 2021.
[16]A. H. Andersen and A. C. Kak, "Simultaneous algebraic reconstruction technique (SART): a superior implementation of the ART algorithm," Ultrasonic imaging, vol. 6, no. 1, pp. 81-94, 1984.
[17]R. Gordon, "A tutorial on ART (algebraic reconstruction techniques)," IEEE Transactions on Nuclear Science, vol. 21, no. 3, pp. 78-93, 1974.
[18]M. M. M. Hossain, G. Lu, and Y. Yan, "Optical fiber imaging based tomographic reconstruction of burner flames," IEEE transactions on instrumentation and measurement, vol. 61, no. 5, pp. 1417-1425, 2012.
[19]Y. Jiang, F. Xu, and B. Xu, "Acoustic emission tomography based on simultaneous algebraic reconstruction technique to visualize the damage source location in Q235B steel plate," Mechanical Systems and Signal Processing, vol. 64, pp. 452-464, 2015.
[20]N. Awasthi, G. Jain, S. K. Kalva, M. Pramanik, and P. K. Yalavarthy, "Deep neural network-based sinogram super-resolution and bandwidth enhancement for limited-data photoacoustic tomography," IEEE transactions on ultrasonics, ferroelectrics, and frequency control, vol. 67, no. 12, pp. 2660-2673, 2020.
[21]J. Jeong, J. Lee, H. Sun, H. Park, S. Kim, and M. S. Bak, "Temperature field estimation of an axisymmetric laminar flame via time-of-arrival measurements of acoustic waves, and machine learning," Experimental Thermal and Fluid Science, vol. 129, p. 110454, 2021.
[22]Q. Kong, G. Jiang, Y. Liu, and J. Sun, "Location of the leakage from a simulated water-cooling wall tube based on acoustic method and an artificial neural network," IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1-18, 2021.
[23]H. Shan, G. Wang, and Y. Yang, "Accelerated correction of reflection artifacts by deep neural networks in photo-acoustic tomography," Applied Sciences, vol. 9, no. 13, p. 2615, 2019.
[24]Y. Wang, H. Guo, T. Feng, J. Ju, and X. Wang, "Acoustic behavior prediction for low-frequency sound quality based on finite element method and artificial neural network," Applied Acoustics, vol. 122, pp. 62-71, 2017.
[25]J. Liu et al., "Applications of deep learning to MRI images: A survey," Big Data Mining and Analytics, vol. 1, no. 1, pp. 1-18, 2018.
[26]S. Pal et al., "An acoustic hotspot tracking algorithm for highly centralized gas temperature distribution," IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 68, no. 4, pp. 1370-1379, 2020.
[27]A. J. Reader, G. Corda, A. Mehranian, C. da Costa-Luis, S. Ellis, and J. A. Schnabel, "Deep learning for PET image reconstruction," IEEE Transactions on Radiation and Plasma Medical Sciences, vol. 5, no. 1, pp. 1-25, 2020.
[28]D. Zankl, S. Schuster, R. Feger, and A. Stelzer, "What a blast!: A massive MIMO radar system for monitoring the surface in steel industry blast furnaces," IEEE Microwave Magazine, vol. 18, no. 6, pp. 52-69, 2017.
[29]M. Bednarik, M. Cervenka, P. Lotton, and G. Penelet, "Behavior of plane waves propagating through a temperature-inhomogeneous region," Journal of Sound and Vibration, vol. 362, pp. 292-304, 2016.
[30]J. Lu, K. Wakai, S. Takahashi, and S. Shimizu, "Acoustic computer tomographic pyrometry for two-dimensional measurement of gases taking into account the effect of refraction of sound wave paths," Measurement Science and Technology, vol. 11, no. 6, p. 692, 2000.
[31]G. Andria, F. Attivissimo, and N. Giaquinto, "Digital signal processing techniques for accurate ultrasonic sensor measurement," Measurement, vol. 30, no. 2, pp. 105-114, 2001.
[32]B. Barshan, "Fast processing techniques for accurate ultrasonic range measurements," Measurement Science and technology, vol. 11, no. 1, p. 45, 2000.
[33]S. Liu, S. Liu, and T. Ren, "Ultrasonic tomography based temperature distribution measurement method," Measurement, vol. 94, pp. 671-679, 2016.
[34]S. Pal et al., "Acoustic Speed Measurement Platform for Monitoring Highly Concentrated Gas Temperature Distribution," IEEE Sensors Letters, vol. 6, no. 3, pp. 1-4, 2022.
[35]R. Raya, A. Frizera, R. Ceres, L. Calderón, and E. Rocon, "Design and evaluation of a fast model-based algorithm for ultrasonic range measurements," Sensors and Actuators A: Physical, vol. 148, no. 1, pp. 335-341, 2008.
[36]S. N. Vecherin, V. E. Ostashev, and D. K. Wilson, "Assessment of systematic measurement errors for acoustic travel-time tomography of the atmosphere," The Journal of the Acoustical Society of America, vol. 134, no. 3, pp. 1802-1813, 2013.
[37]J. Wu, J. Zhu, L. Yang, M. Shen, B. Xue, and Z. Liu, "A highly accurate ultrasonic ranging method based on onset extraction and phase shift detection," Measurement, vol. 47, pp. 433-441, 2014.
[38]T. Motegi, K. Mizutani, and N. Wakatsuki, "Simultaneous measurement of air temperature and humidity based on sound velocity and attenuation using ultrasonic probe," Japanese Journal of Applied Physics, vol. 52, no. 7S, p. 07HC05, 2013.
[39]H. Hernandez, "Standard Maxwell-Boltzmann distribution: definition and properties," ForsChem Research Reports, vol. 2, pp. 2017-2, 2017.
[40]Y. Bao, J. Jia, and N. Polydorides, "Real-time temperature field measurement based on acoustic tomography," Measurement Science and Technology, vol. 28, no. 7, p. 074002, 2017.
[41]M. Barth and A. Raabe, "Acoustic tomographic imaging of temperature and flow fields in air," Measurement Science and Technology, vol. 22, no. 3, p. 035102, 2011.
[42]Y. Li, S. Liu, and S. H. Inaki, "Dynamic reconstruction algorithm of three-dimensional temperature field measurement by acoustic tomography," Sensors, vol. 17, no. 9, p. 2084, 2017.
[43]T. Ma, Y. Liu, and C. Cao, "Neural networks for 3D temperature field reconstruction via acoustic signals," Mechanical systems and signal processing, vol. 126, pp. 392-406, 2019.
[44]X. Glorot, A. Bordes, and Y. Bengio, "Deep sparse rectifier neural networks," in Proceedings of the fourteenth international conference on artificial intelligence and statistics, 2011: JMLR Workshop and Conference Proceedings, pp. 315-323.
[45]V. Nair and G. E. Hinton, "Rectified linear units improve restricted boltzmann machines," in Proceedings of the 27th international conference on machine learning (ICML-10), 2010, pp. 807-814.
[46]S. Y. Chong, J.-R. Lee, and C. Y. Park, "Statistical threshold determination method through noise map generation for two dimensional amplitude and time-of-flight mapping of guided waves," Journal of Sound and Vibration, vol. 332, no. 5, pp. 1252-1264, 2013.
[47]R. Queiros, F. C. Alegria, P. S. Girao, and A. C. C. Serra, "Cross-correlation and sine-fitting techniques for high-resolution ultrasonic ranging," IEEE Transactions on Instrumentation and Measurement, vol. 59, no. 12, pp. 3227-3236, 2010.
[48]F.-S. Lin, M.-C. Huang, S.-K. Tai, Y.-C. Chu, S.-H. Shen, and C.-H. Huang, "Investigation of an FPGA-based acoustic gas thermometer for monitoring indoor temperature distribution two-dimensionally," IEEE Sensors Journal, 2024.