Pig state audio recognition based on underdetermined blind source separation and deep learning

PAN Weihao; SHENG Huizi; WANG Chunyu; YAN Shunpi; ZHOU Xiaobo; GU Lichuan; JIAO Jun

doi:10.7671/j.issn.1001-411X.202312011

Journal of South China Agricultural University > 2024 > 45(5): 730-742. > DOI: 10.7671/j.issn.1001-411X.202312011

PAN Weihao, SHENG Huizi, WANG Chunyu, et al. Pig state audio recognition based on underdetermined blind source separation and deep learning[J]. Journal of South China Agricultural University, 2024, 45(5): 730-742. DOI: 10.7671/j.issn.1001-411X.202312011

Citation:

PDF (3592 KB)

Pig state audio recognition based on underdetermined blind source separation and deep learning

1.
School of Information and Artificial Intelligence, Anhui Agricultural University, Hefei 230036, China
2.
Anhui Xilejia Biotechnology Co., Ltd., Bozhou 233500, China

More Information

Received Date: December 06, 2023
Available Online: June 21, 2024
Published Date: July 14, 2024

Graphical Abstract

Abstract

Abstract

Objective
In order to solve the problem of difficult separation and recognition of pig audio under group rearing environment, we propose a method of pig state audio recognition based on underdetermined blind source separation and ECA-EfficientNetV2.
Method
Four types of pig audio signals were simulated as observation signals in group rearing environment. After the signals were sparsely represented, the signal mixing matrix was estimated by hierarchical clustering, and the l_p-paradigm reconstruction algorithm was used to solve for the minimum of l_p-paradigm to complete the reconstruction of pig audio signals. The reconstructed signals were transformed into acoustic spectrograms, which were divided into four categories, namely, eating sound, roar sound, hum sound and estrous sound. The audio was recognized using the ECA-EfficientNetV2 network model to obtain the state of the pigs.
Result
The normalized mean square error of the hybrid matrix estimation was as low as 3.266×10⁻⁴, and the signal-to-noise ratios of the separated reconstructed audio ranged from 3.254 to 4.267 dB. The acoustic spectrogram was recognized and detected by ECA-EfficientNetV2 with an accuracy of up to 98.35%, and the accuracy improved by 2.88 and 1.81 percentage points compared with the classical convolutional neural networks ResNet50 and VGG16, respectively. Compared with the original EfficientNetV2, the accuracy decreased by 0.52 percentage points, but the amount of the model parameters reduced by 33.56%, the floating-point operations (FLOPs) reduced by 1.86 G, and inference time reduced by 9.40 ms.
Conclusion
The method based on blind source separation and improvement of EfficientNetV2 lightly and efficiently realizes separating and recognizing audio signals of group-raised pigs.
- Pig,
- Blind source separation,
- Spectrogram,
- Audio recognition,
- Sparse reconstruction,
- Convolutional neural network

FullText(HTML)

References (29)

References

[1]	杨亮, 王辉, 陈睿鹏, 等. 智能养猪工厂的研究进展与展望[J]. 华南农业大学学报, 2023, 44(1): 13-23. doi: 10.7671/j.issn.1001-411X.202209050
[2]	GHANI S H, KHAN W. Extraction of UAV sound from a mixture of different sounds[J]. Acoustics Australia, 2020, 48(3): 363-373. doi: 10.1007/s40857-020-00197-z
[3]	HE P, QI M, LI W, et al. A general nonstationary and time-varying mixed signal blind source separation method based on online Gaussian process[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2020, 34(11): 2058015. doi: 10.1142/S021800142058015X
[4]	ADAM A M, FAROUK R M, EL-DESOUKY B S. Generalized gamma distribution for biomedical signals denoising[J]. Signal, Image and Video Processing, 2023, 17(3): 695-704. doi: 10.1007/s11760-022-02277-1
[5]	张振华. 基于音频分析技术的猪异常检测[D]. 太原: 太原理工大学, 2017.
[6]	沈明霞, 王梦雨, 刘龙申, 等. 基于深度神经网络的猪咳嗽声识别方法[J]. 农业机械学报, 2022, 53(5): 257-266. doi: 10.6041/j.issn.1000-1298.2022.05.026
[7]	JI N, SHEN W Z, YIN Y L, et al. Investigation of acoustic and visual features for pig cough classification[J]. Biosystems Engineering, 2022, 219: 281-293.
[8]	彭硕, 陶亮, 查文文, 等. 基于稀疏分量分析的生猪音频欠定盲源分离研究[J]. 畜牧兽医学报, 2023, 54(7): 2794-2809.
[9]	HU Y, LOIZOU P C. Speech enhancement based on wavelet thresholding the multitaper spectrum[J]. IEEE Transactions on Speech and Audio Processing, 2004, 12(1): 59-67. doi: 10.1109/TSA.2003.819949
[10]	杨稷, 沈明霞, 刘龙申, 等. 基于音频技术的肉鸡采食量检测方法研究[J]. 华南农业大学学报, 2018, 39(5): 118-124. doi: 10.7671/j.issn.1001-411X.2018.05.017
[11]	赵欢, 王纲金, 赵丽霞. 一种新的对数能量谱熵语音端点检测方法[J]. 湖南大学学报(自然科学版), 2010, 37(7): 72-77.
[12]	朱国俊, 唐振博, 冯建军, 等. 启动方式对混流泵噪声特性的影响[J]. 农业工程学报, 2023, 39(13): 34-42. doi: 10.11975/j.issn.1002-6819.202304091
[13]	杜晓冬, 滕光辉, 刘慕霖, 等. 基于轻量级卷积神经网络的种鸡发声识别方法[J]. 农业机械学报, 2022, 53(10): 271-276. doi: 10.6041/j.issn.1000-1298.2022.10.029
[14]	CAI W, LI M. A unified deep speaker embedding framework for mixed-bandwidth speech data [EB/OL]. arXiv: 2012.00486[2023-12-01]. https://ui.adsabs.harvard.edu/abs/2020arXiv201200486C.
[15]	丁楠. 基于特征学习的语音情感识别研究[D]. 南京: 南京邮电大学, 2023.
[16]	CANDAN C. An automated window selection procedure For DFT based detection schemes to reduce windowing SNR loss[EB/OL]. arXiv: 1710.10200[2023-12-01]. https://ui.adsabs.harvard.edu/abs/2017arXiv171010200C.
[17]	张逸, 陈书畅, 刘必杰, 等. 基于盲源分离的工业谐波源负荷分类识别方法[J]. 中国电机工程学报, 2024, 44(10): 3850-3862.
[18]	邹亮, 张鹏, 陈勋. 基于三阶统计量的欠定盲源分离方法[J]. 电子与信息学报, 2022, 44(11): 3960-3966. doi: 10.11999/JEIT210844
[19]	凌康杰, 岳学军, 刘永鑫, 等. 基于移动互联的农产品二维码溯源系统设计[J]. 华南农业大学学报, 2017, 38(3): 118-124. doi: 10.7671/j.issn.1001-411X.2017.03.019
[20]	王晶, 李炜, 洪心睿, 等. 基于改进密度聚类算法的语音信号欠定盲分离[J]. 信息与控制, 2023, 52(6): 784-796.
[21]	KEMIHA M, KACHA A. Single-channel blind source separation using adaptive mode separation-based wavelet transform and density-based clustering with sparse reconstruction[J]. Circuits Systems and Signal Processing, 2023, 42(9): 5338-5357. doi: 10.1007/s00034-023-02350-1
[22]	WANG M, CAI X X, ZHU K F. Underdetermined mixed matrix estimation of single source point detection based on noise threshold eigenvalue decomposition[C]//Proceedings of the 8th International Conference on Communications, Signal Processing, and Systems. Singapore: Springer, 2020: 704-711.
[23]	LI Y B, NIE W, YE F, et al. A complex mixing matrix estimation algorithm in under-determined blind source separation problems[J]. Signal, Image and Video Processing, 2017, 11(2): 301-308. doi: 10.1007/s11760-016-0937-y
[24]	冯建英, 石岩, 王博, 等. 基于聚类分析的数据挖掘技术及其农业应用研究进展[J]. 农业机械学报, 2022, 53(S1): 201-212. doi: 10.6041/j.issn.1000-1298.2022.S1.021
[25]	XIE Y, XIE K, XIE S L. Underdetermined blind separation of source using l _p-norm diversity measures[J]. Neurocomputing, 2020, 411: 259-267.
[26]	XU P F, JIA Y J, JIANG M X. Blind audio source separation based on a new system model and the Savitzky-Golay filter[J]. Journal of Electrical Engineering, 2021, 72(3): 208-212. doi: 10.2478/jee-2021-0029
[27]	FISLI S, DJENDI M. Hybrid PSO-NLMS (HPSO-NLMS) algorithm for blind speech quality enhancement in time domain[J]. Applied Acoustics, 2021, 177: 107936.
[28]	TAN M, LE Q V. EfficientNetV2: Smaller models and faster training[EB/OL]. arXiv: 2104.00298[2023-12-01]. https://ui.adsabs.harvard.edu/abs/2021arXiv210400298T.
[29]	WANG X K, JIA X, ZHANG M Y, et al. Object detection in 3D point cloud based on ECA mechanism[J]. Journal of Circuits, Systems and Computers, 2023, 32(5): 2350080.

Cited By

Cited by

Periodical cited type(6)

1.	高旻. 视频分析技术在港口交通流量监控中的应用. 电视技术. 2025(01): 223-225 .
2.	陈钇果. 基于FPGA架构的人工智能加速器设计与实现. 软件. 2025(01): 59-61 .
3.	沃英达. 基于深度学习算法的激光散斑音频信号提取与增强. 电声技术. 2025(02): 52-54+77 .
4.	祁昊，裴杨莉，向海，李川，朱晓萍. 以畜牧学基础课程教学为例探讨课程思政的教学改革. 当代畜牧. 2025(03): 92-95 .
5.	蓝壮青. 基于欠定盲源分离的双路音频信号噪声自适应分离. 现代电子技术. 2024(24): 68-72 .
6.	吴源源. 基于机器学习的音频信号处理系统设计. 电声技术. 2024(11): 182-184 .

Other cited types(0)

Get Citation

PDF

XML

Article views (388) PDF downloads (24) Cited by(6)

Turn off MathJax

Article Contents

Abstract

References

Pig state audio recognition based on underdetermined blind source separation and deep learning

Abstract

References

Cited by

Periodical cited type(6)

Other cited types(0)

Catalog

Export File

Citation

Format

Content