EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Adaptive Decomposition Level Selection for Wavelet-Based Dolphin Whistle Denoising Using PDF Similarity

Jiali Chen, Peibin Zhu, Zhong Liu, Wen Chen, Xiaomei Xu, Yougan Chen

IEEE Access · 2026 · 14 · 28739-28753

DOI: 10.1109/access.2026.3666548 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Jiali Chen und 5 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Science Foundation of Fujian Science and Technology Plan, Xiamen Science and Technology Subsidy Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In passive acoustic monitoring, effective noise reduction of dolphin whistles is frequently compromised by ambient noise. A key challenge for wavelet-based denoising techniques lies in the selection of decomposition levels, where inappropriate choices often result in signal feature loss or residual noise. To address this issue, this paper proposes an adaptive method based on noise probability density function (PDF) similarity to determine the optimal decomposition level. This approach automatically selects the optimal decomposition level by comparing the probability density functions of the removed noise across different decomposition levels and quantifying the differences between adjacent levels using the total variance distance. This eliminates the need for prior information, thereby enhancing both the efficiency and accuracy of noise reduction. The effectiveness of the proposed method was validated through experiments on simulated dolphin whistle signals under various noise conditions and on real dolphin whistle signals. Experimental results demonstrate that the PDF-based approach consistently and accurately identifies the optimal decomposition level, enabling subsequent denoising stages to effectively eliminate noise while preserving the key structural features of dolphin whistle signals. This method provides a robust automated solution for processing vast amounts of raw data from passive acoustic monitoring systems, ultimately enhancing the precision of bioacoustic analysis and which supports marine ecosystem conservation efforts.

Förderung

  • Natural Science Foundation of Fujian Science and Technology Plan
  • Xiamen Science and Technology Subsidy Project

Zitieren

APA

Chen, J., Zhu, P., Liu, Z., Chen, W., Xu, X., & Chen, Y. (2026). Adaptive Decomposition Level Selection for Wavelet-Based Dolphin Whistle Denoising Using PDF Similarity. IEEE Access, 14, 28739-28753. https://doi.org/10.1109/access.2026.3666548

Vancouver

Chen J, Zhu P, Liu Z, Chen W, Xu X, Chen Y. Adaptive Decomposition Level Selection for Wavelet-Based Dolphin Whistle Denoising Using PDF Similarity. IEEE Access. 2026;14:28739-28753. doi:10.1109/access.2026.3666548

BibTeX
@article{Chen2026,
  author = {Chen, Jiali and Zhu, Peibin and Liu, Zhong and Chen, Wen and Xu, Xiaomei and Chen, Yougan},
  title = {Adaptive Decomposition Level Selection for Wavelet-Based Dolphin Whistle Denoising Using PDF Similarity},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {28739--28753},
  doi = {10.1109/access.2026.3666548}
}
Kostenlos starten →