Automatic monitoring of grouphoused pigs in real time through porcine acoustic signals has played a crucial role in automated farming. In the process of data collection and transmission, acoustic signals are generally interfered with noise. In this paper, an effective porcine acoustic signal denoising technique based on ensemble empirical mode decomposition (EEMD), independent component analysis (ICA), and wavelet threshold denoising (WTD) is proposed. Firstly, the porcine acoustic signal is decomposed into intrinsic mode functions (IMFs) by EEMD. In addition, permutation entropy (PE) is adopted to distinguish noisedominant IMFs from the IMFs. Secondly, ICA is employed to extract the independent components (ICs) of the noisedominant IMFs. The correlation coefficients of ICs and the first IMF are calculated to recognize noise ICs. The noise ICs will be removed. Then, WTD is applied to the other ICs. Finally, the porcine acoustic signal is reconstructed by the processed components. Experimental results show that the proposed method can effectively improve the denoising performance of porcine acoustic signal.
With the development of precision livestock farming, it is hard for breeders to monitor porcine abnormal states. Sound recognition, as one of the noncontact detection methods, has been proven to be a valuable method to detect some diseases [
Empirical mode decomposition (EMD) is an effective automatic decomposition algorithm to analyze nonlinear, nonstationary, and nonGaussian signals [
In order to effectively eliminate the noise produced in the process of sound collection and transmission, EEMDICAWTD, which can be employed before porcine sound recognition, is proposed in this paper. EEMD is used to decompose the porcine acoustic signal into IMFs. Then noisedominant IMFs are distinguished by permutation entropy (PE). The independent components (IC) of noisedominant IMFs are extracted by independent component analysis (ICA). As the first IMF contains much of the highfrequency noise [
This paper is organized as follows: Section
The materials of this study are porcine acoustic signals. Below are the details of porcine acoustic signals. The methods are mainly comprised of EEMD, PE, FastICA, and WTD, with a detailed explanation given below.
In this study, the original data are collected by an acoustic pickup device (ELITE model OS100) from a largescale pig farm in Shanxi Province, China. The schematic of the installation of the acoustic pickup in the pig farm is shown in Figure
The installation drawing of the acoustic pickup in piggery.
The process of EEMDICAWTD is mainly comprised of porcine acoustic signal decomposition based on EEMD, noisedominant IMFs differentiation based on PE, independent components extraction by FastICA, and denoising by WTD.
In this paper, the porcine acoustic signal is firstly decomposed. EMD can decompose the signals into IMFs from high to low frequency selfadaptively [
The local extreme points which are detected from the original signal
The list
The residual list
The residual list
Therefore, the original signal
Classical EMD may cause frequency aliasing during signal decomposition. In order to overcome this shortcoming, EEMD was proposed by Wu and Huang [
Add a random Gaussian white noise signal
where
Decompose
where
Repeat the process as described above
where
Permutation entropy (PE), proposed by Bandt and Pompe [
For a given time series
Each row of
Since the embedding dimension is
The PE of order can be normalized as [
The ICA, as one of the multivariate statistical methods, is widely used in statistical sources separation [
The independent sources can be denoted as
This function can be expressed by the matrix:
The aim of ICA is to estimate the inverse of the mixing matrix
FastICA algorithm is one of the improved ICA algorithms, which is widely utilized to estimate the orthogonal matrix. FastICA has higher convergence speed compared to the conventional method and the stepsize parameters are not needed. In this paper, FastICA is used to extract independent components from IMFs.
Wavelet transform denoising (WTD) is one of the denoising algorithms based on wavelet transform (WT). WT can decompose signals at different scales. The discrete wavelet transform (DWT) is calculated as follows:
The primary steps of WTD are described as follows:
Decompose the original signal by WT with proper wavelet basis function and decomposition level.
The threshold is performed by the selected proper threshold method for highfrequency coefficients at each decomposition scale. The lowfrequency wavelet coefficient is kept unchanged.
The signal is reconstructed by the lowfrequency coefficients and highfrequency coefficients after threshold processing.
It is crucial to select an appropriate threshold method for WTD. The common threshold selection methods fall into soft threshold method and hard threshold method [
The process of the new efficient denoising technique proposed in this paper is shown in Figure
The porcine acoustic signal is decomposed into IMFs by EEMD. Sorted in the increasing order of IMFs, the frequency distribution of the IMFs varies from high to low. The noise mainly concentrates in high frequency. Therefore, the first few IMFs contain both the information of porcine acoustic signal and noise [
Denoising the noisedominant IMFs directly may destroy the continuity of reconstructed signals. It is harmful to the denoising effect [
As the first IMF contains much of the highfrequency noise [
Denoise the other ICs by WTD. The wavelet basis function and decomposition level we selected are db6 and 3.
The denoised ICs (DICs) are transformed to denoised IMFs (DIMFs) through the matrix of mixing coefficients. Then the porcine acoustic signal is reconstructed by these DIMFs and real IMFs.
The process chart of the proposed denoising technique for porcine acoustic signal.
This section introduces the simulation process and results of EEMDICAWTD. In order to verify the performance of EEMDICAWTD, the performance of the denoising is compared with the other six methods.
In order to analysis the process and results of EEMDICAWTD, porcine scream and porcine cough are selected for denoising, taking the noisy scream with 5 dB SNRin by adding Gaussian white noise as an example. The timedomain waveforms of porcine scream and noisy scream are shown in Figure
The timedomain waveforms of (a) porcine scream and (b) noisy scream.
The noisy scream is decomposed as step 1. The timedomain waveforms of IMFs and Res are shown in Figure
The decomposition result of the noisy scream.
The PE of each IMF is calculated as step 1. The time delay
PEs of noisy scream IMFs.
IMF1  IMF2  IMF3  IMF4  IMF5  IMF6  IMF7  IMF8  IMF9  Res 

0.974  0.997  0.824  0.674  0.567  0.493  0.440  0.415  0.396  0 
Table
The ICs of noisedominant IMFs are extracted by FastICA as step 2. The timedomain waveforms of ICs are shown in Figure
The timedomain waveforms of ICs.
The correlation coefficients are calculated as step 3. The correlation coefficients of ICs and the first IMF are shown in Table
Correlation coefficients of each IC and the first IMF.
IC1  IC2  IC3  IC4  IC5 

0.3222  0.4656  0.2772  0.8460  0.3170 
Table
The other ICs are denoised by WTD as step 4. The denoising results are shown in Figure
The denoising results of ICs by WTD.
The end result of the reconstructed signal is shown in Figure
The timedomain waveforms of the reconstructed porcine scream.
In order to evaluate the denoising performance of the method quantitatively, the root mean square error (RMSE), SNRout, and correlation coefficient (
RMSE reflects the degree of error between the denoised porcine acoustic signal and the original porcine acoustic signal. The smaller the value, the better the denoising effect. SNRout reflects the ratio of the porcine acoustic signal to real noise. Therefore, the higher the value, the less noise mixes.
The performance of the denoising is shown in Table
The denoising parameters of porcine scream.
RMSE  SNRout 


0.1476  6.2831  0.8992 
In order to verify the performance of EEMDICAWTD, six different denoising methods are used as comparison methods. They are EMDTD [
Denoising results of porcine scream.
SNRin  Parameter  Denoising methods  

EMDTD  EMDWTD  EEMDTD  EEMDWTD  WSTD  MBSS  EEMDICAWTD  
−10  RMSE  0.4794  0.3594  0.3952  0.3682  0.4319  0.7362  0.3389 
SNRout/dB  −3.9501  −1.4474  −2.2729  −1.6579  −3.0440  −8.6582  −1.0408  

0.0633  0.3585  0.1723  0.3303  0.1762  0.0324  0.3701  


−5  RMSE  0.4379  0.3597  0.3774  0.3643  0.4101  0.5214  0.3339 
SNRout/dB  −3.1633  −1.2094  −1.8728  −1.5668  −2.5937  −3.9452  −0.8091  

0.1449  0.5380  0.2498  0.5343  0.3897  0.1167  0.5544  


0  RMSE  0.4497  0.2533  0.3389  0.2601  0.2867  0.3241  0.2496 
SNRout/dB  −3.3946  1.5894  −0.9377  1.3619  0.5148  −0.8852  1.4573  

0.1972  0.7590  0.4029  0.7779  0.7683  0.7091  0.7810  


5  RMSE  0.4460  0.1474  0.3483  0.1491  0.1506  0.1543  0.1432 
SNRout/dB  −3.3242  6.2958  −1.1767  6.1937  6.1054  6.2438  6.4831  

0.2381  0.9047  0.3955  0.9064  0.9029  0.8823  0.9145  


10  RMSE  0.4315  0.0971  0.3321  0.0944  0.0935  0.0951  0.0913 
SNRout/dB  −3.0361  9.9147  −0.7613  10.1594  10.2463  10.0864  10.3106  

0.2038  0.9480  0.4088  0.9518  0.9531  0.9487  0.9569 
Denoising results of porcine cough.
SNRin  Parameter  Denoising methods  

EMDTD  EMDWTD  EEMDTD  EEMDWTD  WSTD  MBSS  EEMDICAWTD  
−10  RMSE  0.4216  0.3100  0.3245  0.3163  0.3417  0.4775  0.3025 
SNRout/dB  −2.4745  0.1955  −0.1995  0.0217  −0.6479  −2.9245  0.2213  

0.2337  0.4732  0.4832  0.4951  0.4389  0.2183  0.4978  


−5  RMSE  0.3447  0.2629  0.2666  0.2773  0.2946  0.3341  0.2442 
SNRout/dB  −0.7254  1.6290  1.5076  1.1645  0.6390  −0.4762  1.9498  

0.3903  0.6406  0.6656  0.6736  0.6075  0.4097  0.6752  


0  RMSE  0.4168  0.2048  0.2464  0.2042  0.2423  0.2658  0.1880 
SNRout/dB  −2.3755  3.7967  0.1914  3.8236  2.3360  2.1935  4.0920  

0.2767  0.7936  0.7359  0.8310  0.8257  0.7016  0.8353  


5  RMSE  0.3941  0.1298  0.2762  0.1343  0.1362  0.1407  0.1224 
SNRout/dB  −1.8887  7.7609  1.1993  7.4601  7.3379  7.2837  7.9737  

0.2712  0.9237  0.7403  0.9332  0.9293  0.8819  0.9360  


10  RMSE  0.2777  0.0721  0.0952  0.0692  0.0842  0.0743  0.0646 
SNRout/dB  2.1711  13.8798  11.4739  14.2417  12.5417  13.7961  14.6788  

0.6897  0.9805  0.9689  0.9854  0.9828  0.9829  0.9859 
The results, shown in Tables
The results show that the EEMDICAWTD proposed in this paper has the best denoising effects with different SNRins not only for porcine scream but also for porcine cough. The EEMDWTD has the secondbest denoising effects. Taking the denoising results for porcine cough as an example, when the SNRin of porcine cough is 10 dB, the values of RMSE, SNRout, and
Denoising results for porcine scream with different SNRins: (a) RMSE, (b) SNRout, and (c)
Denoising results for porcine cough with different SNRins: (a) RMSE, (b) SNRout, and (c)
To improve the denoising performance of porcine acoustic signal, an efficient denoising technique based on EEMDICAWTD is proposed in this paper. The approach has been developed with the purpose to reduce noise interference during the recognition of porcine abnormal sounds.
Firstly, the porcine acoustic signal is decomposed into different components in order of frequency. Because of the frequency aliasing of EMD, the EEMD is used to decompose the porcine acoustic signal into IMFs. As the noise mainly concentrates in high frequency, PE is used to distinguish the noisedominant IMFs from the IMFs. Secondly, the continuity of the signal may be adversely affected if the noisedominant IMFs are denoised directly. Therefore, the ICs of noisedominant IMFs are extracted by FastICA. The noise and real information are concentrated on the ICs. It has been shown that the first IMF contains much highfrequency noise. Therefore, the noise ICs are identified by correlation coefficients of ICs and the first IMF and are then removed. Finally, WTD is used for denoising the other ICs. The porcine acoustic signal is then reconstructed by processed ICs. The performance of this denoising method is shown to be superior to other methods.
In the future work, this approach will be optimized to reduce the run time on the premise of guaranteeing the performance of the denoising.
The porcine acoustic data used to support the findings of this study are included within the supplementary information file.
The authors declare that there are no conflicts of interest regarding the publication of this paper.
This study was supported by the National High Technology Research and Development Program of China (863 Program) (2013AA102306).
The supplementary materials are porcine acoustic signals of scream and cough. The data are saved as WAV format. The length of each signal is 1 s.