Accurate TOF measurement of ultrasonic signal echo from the liquid level based on a 2-D image processing method
Introduction
Range detection is used in many instrumental and industrial applications. Liquid level measurement is a type of range detection. For liquid level measurement, a variety of principles have been proposed, including mechanical, optical, electromagnetic and acoustic methods [1]. The low-cost Microsoft Kinect sensor has been applied to find the highly accurate depth information to solve some fundamental problems in computer vision. Therefore, the Microsoft Kinect sensor can be used in liquid level measurement [2]. Currently, however, the ultrasonic method is still widely applied to liquid level measurement because of its cost effectiveness. In the past decade, various ultrasonic level measurements have been developed to achieve high reliability and accuracy. However, the conventional method is still the main technique. The so-called conventional method is a technique that transmits a sine pulse and then receives the echo signal from the liquid level using a single-channel system. The key of the conventional level measurement method is to accurately estimate an important parameter, the time of flight (TOF), from the received echo signal [3], [4], [5]. The measurement accuracy of the TOF is directly related to the accuracy of the level measurement.
The TOF measurement is traditionally performed by simple threshold detection; that is, a constant threshold is set for the received echo signal wave shape and the TOF is determined by locating the first wave peak greater than the threshold [6]. An improved method is to calculate the slope of the rising edge of the echo wave shape to estimate TOF [7]. However, if the signal-to-noise ratio (SNR) is poor, these methods may not be effective [8]. The cross-correlation method can reduce the dependence on the high SNR of the echo signal [9]. Based on the maximum likelihood criterion, the cross-correlation estimator can provide the optimal solution if the signal is undistorted. However, the frequency-dependent scattering and attenuation in the process of propagation may distort the echo signal [10], and the TOF estimates with the cross-correlation method may have larger measuring errors [9]. In addition, this method requires a reference, signal which may not be available.
Model-based TOF measurements can overcome some of the shortcomings of the cross-correlation method because they do not need reference signals and can obtain high precision TOF estimates. However, these model-based methods have less reliability when multiple echoes exist in the received signal, a mass of parameters needs to be estimated by solving a non-linear optimization problem. For this situation, reducing the dimensions of the problem through transformation is often desired. By considering only the envelope of the echo signals, the number of parameters can be reduced, and the accuracy of TOF measurement can be improved [11]. The least squares method is used to estimate the parameters of the envelope of the ultrasonic echo signal [11]. However, the least squares method is optimal only when the noise influencing the echo signal is white Gaussian noise (WGN). The quasi-maximum likelihood method is also used to estimate the TOF [12]. These modified model-based methods can improve parameter estimation, which ultimately enhances detection and assessment.
In an actual level measurement using the sine pulse as the transmitted signal, the pulse boundary of the ultrasonic echo signal approximates an elliptical shape in the time–amplitude drawing due to the energy attenuation and absorption in liquid [13], [14]. Therefore, ellipse boundary fitting is employed to estimate the TOF, in which the time corresponding to the vertical axis of the ellipse is regarded as the TOF. However, the real TOF is a certain time between the starting time of the ellipse pulse and that corresponding to the long axis [15]. In this paper, we propose a measurement method of TOF that combines the ellipse fitting algorithm and the Canny edge detector in 2-D image processing. Following extraction of the oval echo pulse with the ellipse fitting algorithm, the boundary of the ellipse is detected by the Canny edge operator, and a differential equation is solved to search the maximum gradient of the extracted ellipse boundary in the second and third quadrants. The time corresponding to the largest gradient is considered as the TOF. Thus, a more accurate TOF can be obtained because the boundary characteristics of the echo signal are fully used.
This paper is organized as follows. Section 2 describes two TOF measurement methods, including the threshold method and the central ellipse algorithm. On the basis of the existing ellipse central algorithm, a modified method is proposed in Section 3. In Section 4, we collect the echo signals from the real liquid levels to verify the modified method. Then, the relevant results are analysed and evaluated to confirm the effectiveness of the method. In Section 5, several possible influencing factors for the modified method are analysed. Section 6 concludes the paper.
Section snippets
Central ellipse algorithm
For liquid level measurement and other range detections, a sine pulse is often used as the transmitted signal, which can be expressed aswhere is the amplitude of the transmitted signal, is the carrier frequency, and is the initial phase. Assuming that the TOF is and is the attenuation function of the sound wave, the echo signal received by the ultrasonic transducer can be described as
The waveforms of the transmitted pulse
Modified ellipse algorithm
According to the theory of sound wave propagation, part of a sound wave will be reflected when the sound wave encounters the interface of two layers of different media [19]. A sudden change of acoustic pressure occurs when the sound wave is reflected. Correspondingly, the amplitude of the received echo pulse signal also changes suddenly. Therefore, we suppose that the TOF is the time corresponding to the sudden change in the amplitude of the echo pulse. Under this assumption, we propose a
Real echo data acquisition
To verify the effectiveness and accuracy of the modified TOF method, we build a real data acquisition system to collect the ultrasonic echo signals from the liquid level. The system parameters are set as follows: duration of transmitted pulse, 0.08 ms;centre frequency, 150 KHz; sampling frequency, 2 MHz; temperature, 20 °C; and corresponding velocity of sound, 343.2 m/s. Liquid level detections with different ranges from the transducer to the liquid level are implemented to verify the influence of
Analysis of influential factors
The experimental results demonstrate that the modified method based on the central ellipse algorithm significantly improves the TOF measuring accuracy when using the ultrasonic liquid level measurement approach. To further verify the reliability of the method, we also examine related influencing factors.
Conclusions
In this paper, we improve the existing central ellipse algorithm to enhance the estimation accuracy of TOF of conventional ultrasonic level measurement. The difference between the two methods is that the modified method uses a more reasonable approach than the vertical axis of the extracted ellipse to estimate TOF. In the modified method, the Canny edge detector is used to extract the boundary of the binary image of the echo pulse signal after the existing central ellipse algorithm, and the
Acknowledgments
This work was supported by a Grant from the National Natural Science Foundation of China (No. 41075115), the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (No. 10KJB510012) and the Jiangsu Key R&D Program (BE2015692). This work is also part of a project funded by the Priority Academic Program Development (PAPD) of Jiangsu Higher Education Institutions. In addition, the author would like to thank the anonymous referee for invaluable comments and suggestions.
Peng Li received the B.S. degree in electronic engineering from the Nanjing University of Science and Technology, Nanjing, China, in 1990, the M.S. degree in electronic science and technology from the China University of Mining and Technology, Xuzhou, China, in 2003, and the Ph.D. degree in biomedical engineering from Xi׳an Jiaotong University, Xi׳an, China, in 2008. He became a Senior Engineer in 2002. He has been involved in medical instrument-related technology and research since 1990. He is
References (23)
- et al.
Shear-wave time of flight diffraction (S-TOFD) technique
NDT E Int.
(2006) - et al.
Estimating ultrasonic time of flight using envelope and quasi maximum likelihood method for damage detection and assessment
Measurement
(2012) - et al.
Non-contact liquid level measurement with electromagnetic acoustic resonator sensors
Meas. Sci. Technol.
(2009) - et al.
Enhanced computer vision with Microsoft Kinect sensor: a review, IEEE Trans
Cybern
(2013) - et al.
Digital time-of-flight measurement for ultrasonic sensors
IEEE Trans. Instrum. Meas.
(1992) - et al.
A successive parameter estimation algorithm for chirplet signal decomposition
IEEE Trans. Ultrason. Ferroelectr. Freq. Control
(2006) - R. Jansohn, M. Schickert. Objective interpretation of ultrasonic concrete images, in: Proceedings of the 7th European...
- et al.
Estimating ultrasonic time-of-flight through quadrature demodulation
IEEE Trans. Instrum. Meas.
(2006) - et al.
Measurement of ultrasonic velocity using phase-slope and cross-correlation method
Mater. Eval.
(1985) - et al.
A new Golay code system for ultrasonic pulse echo measurements
Meas. Sci. Technol.
(1990)
Model-based estimation of ultrasonic echoes. Part I: analysis and algorithms
IEEE Trans. Ultrason. Ferroelectr. Freq. Control
Cited by (19)
Investigation on high temperature point detection of spontaneous combustion of loose coal based on optimal acoustic signal
2024, Process Safety and Environmental ProtectionAccuracy-improved ultrasonic phase algorithm for measuring lubricant film thickness
2024, International Journal of Mechanical SciencesLiquid-level response measurement using high-speed videogrammetry with robust multiple sphere tracking
2024, Measurement: Journal of the International Measurement ConfederationA sensitivity-improved amplitude method for determining film thickness based on the partial reflection waves
2023, Tribology InternationalICD: A methodology for real time onset detection of overlapped acoustic emission waves
2020, Automation in ConstructionCitation Excerpt :In the literature many identification methods have been proposed such as using thresholding [12], Short term and Long Term Average (STA/LTA) [13], Akaike information criteria (AIC) [8,14], using cross-correlation from a binary image [12,15] and wavelet analysis and Hilbert transform [16]. Advanced TOA detection methods that are popular among (ultrasonic) practitioners includes traditional quadrature-demodulation scheme [17], synthetic aperture focusing technique (SAFT) along with constructive interference [18], square root unscented Kalman filter (SRUKF) method [19] improved self-interference driving technique [20], correlating input and output to high signal-to-noise ratio waveforms then using L2-norm [21] and using the canny operator and modified ellipse algorithm [22]. These TOA detection methods, as well as conventional time-frequency decomposition methods Short-time Fourier, transform (STFT), Wigner–Ville transforms (WVT), Wavelet Transform (WT) and the Empirical Mode Decomposition (EMD) perform poorly (not only when the waves are overlapped but even) when waveforms are relatively close in time and frequency [23].
Slag adhesion on the refractory sensor for molten steel level measurement
2018, International Journal of Heat and Mass TransferCitation Excerpt :The molten steel is up to 1500 °C, disabling the applications of contact measuring methods, i.e. pressure gauges [3,4], fiber laser sensors [5], etc. The covering slag of unknown thickness and compositions floats on top of the molten steel, limiting the usage of non-contact measuring techniques such as ultrasonic waves [6,7], optical measurement [8,9] and microwaves [10,11]. Although radioactive gamma rays [12] have been commonly used for molten steel level measurement in the mold, they are not suitable for open places because of high energy radiation.
Peng Li received the B.S. degree in electronic engineering from the Nanjing University of Science and Technology, Nanjing, China, in 1990, the M.S. degree in electronic science and technology from the China University of Mining and Technology, Xuzhou, China, in 2003, and the Ph.D. degree in biomedical engineering from Xi׳an Jiaotong University, Xi׳an, China, in 2008. He became a Senior Engineer in 2002. He has been involved in medical instrument-related technology and research since 1990. He is teaching Sensors and Measurement Instrument with the Department of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing, where he is currently an Associate Professor. His current research interests include measurement of weather conditions, meteorological sensors and instruments, and acoustics imaging and related signal processing.
Sai Chen received the B.S. degree from Nanjing University of Information Science and Technology, Nanjing, China, in 2013. He is currently pursuing the M.E. degree with the Department of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing, China. His main research interests are signal and information processing and Ultrasound medical imaging.
Yulei Cai received the B.S. degree from Ludong University, Yantai, China, in 2012, and received the M.E. degree with the Department of Electronic and Information Engineering from Nanjing University of Information Science and Technology, Nanjing, China, in 2015. His current research interests include signal and information processing and MIMO ultrasonic imaging system.
Jiaqiang Li received the B.S. and M.S. degrees in resources and information science from the China University of Petroleum, Qingdao, China, in 2000 and 2003, respectively, and the Ph.D. degree in electronic and electrical engineering from Shanghai Jiaotong University, Shanghai, China, in 2007. His current research interests include radar systems and signal processing.
Jinli Chen received the Ph.D. degree from Nanjing University of Science & Technology, Nanjing, China, in 2010. Now, he is a Lecturer in College of Electronic & Information Engineering, Nanjing University of Information Science & Technology, Nanjing, China. His research is focused on array signal processing, communication signal processing, and MIMO radar array signal processing.