生物多样性声学指数的间距互质子阵融合降噪方法

陈蕾; 许志勇; 赵兆

doi:10.11999/JEIT260237

生物多样性声学指数的间距互质子阵融合降噪方法

doi: 10.11999/JEIT260237 cstr: 32379.14.JEIT260237

陈蕾¹,
许志勇^1, ,,
赵兆¹

南京理工大学南京 210094

详细信息

作者简介:
陈蕾：女，博士研究生，研究方向为阵列信号处理

许志勇：男，副教授，研究方向为大气声探测系统技术、阵列信号处理、生物声学分析与处理

赵兆：男，副教授，研究方向为麦克风阵列信号处理、智能信号处理与大气被动声探测系统设计与实现

通讯作者:
许志勇　ezyxu@njust.edu.cn

中图分类号: TN911
计量
- 文章访问数: 15
- HTML全文浏览量: 11
- PDF下载量: 1
- 被引次数: 0
出版历程
- 收稿日期: 2026-03-05
- 修回日期: 2026-05-15
- 录用日期: 2026-05-15
- 网络出版日期: 2026-06-03

A Noise Reduction Strategy via Coprime-Spacing Subarrays for Biodiversity Acoustic Indices

CHEN Lei¹,
XU Zhiyong^{1
, ,},
ZHAO Zhao¹

School of Electronic and Optical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China

摘要

摘要: 针对野外生态监测录音数据中包括人为干扰声在内的背景噪声导致生物多样性声学指数失真问题，利用声学指数典型应用场景中人为干扰声通常具有的空域点声源和时域多线谱特征，结合对麦克风阵列空时二维频响起伏不敏感的依频声学多样性指数，提出一种基于间距互质子阵融合的空时自适应降噪方法，进而形成一种新的噪声鲁棒声学指数——自适应干扰对消依频声学多样性指数(AIC-FADI)。该方法采用三个麦克风构建两个间距较宽且互质的双元子阵组成非均匀线阵，两个子阵分别进行空时自适应白化滤波滤除平稳定向干扰，然后利用宽间距阵列的高空间分辨性质及间距互质子阵空时二维频响之间的空域混叠零陷交错特性，通过对两者输出信号的二值化时频谱进行逐点取大互补融合，重构生物声信号在干扰方向二维零陷之外的二值化时频分布结构特征，缓解无约束干扰对消导致的目标自消问题，使所提AIC-FADI的计算结果能在抑制定向干扰及其它背景噪声影响的同时，尽可能完整保留全空域生物声学信息的统计特征，从而推动声学指数的实际应用场景和时空覆盖范围实现实质性拓展。
- 麦克风阵列 /
- 自适应干扰抑制 /
- 空时自适应滤波 /
- 空域混叠影响抑制 /
- 声学指数
Abstract: Objective As a popular tool for rapid biodiversity assessment, acoustic indices have attracted increasing attention in the field of soundscape ecology in recent years. Nevertheless, most commonly used acoustic indices are susceptible to background noise. Traditional single-channel noise reduction strategies, including spectral subtraction, high-pass filtering, and threshold detection, have been widely adopted as preprocessing approaches to optimize the calculation of acoustic indices. However, when dealing with anthropogenic interference that overlaps with biotic signals in both time and frequency domains, the denoising capability of single-channel methods degrades severely. Although spatio-temporal adaptive whitening filtering based on microphone arrays provides a feasible approach for suppressing directional interference, it suffers from a non-uniform two-dimensional spatio-temporal amplitude response and the self-cancellation of target signal in the unconstrained interference cancellation. These disadvantages lead to distortion in the time-frequency distribution of target signals, causing acoustic index calculations to deviate from the ground truth. Therefore, this study aims to propose a noise reduction strategy via coprime-spacing subarrays for biodiversity acoustic indices. This method effectively suppresses directional interference while maximally preserving the time-frequency distribution structure of biotic signals. Methods The noise reduction strategy based on microphone array spatio-temporal adaptive whitening filtering is proposed, incorporating the Frequency-dependent Acoustic Diversity Index (FADI), which is insensitive to fluctuations in the array's two-dimensional spatio-temporal amplitude response. A noise-robust acoustic index method, termed Adaptive Interference Cancellation–Frequency-dependent Acoustic Diversity Index (AIC-FADI), is subsequently developed. Specifically, a non-uniform linear array is first constructed using three microphones to form two dual-element subarrays with coprime spacing. This design fully exploits the high spatial resolution of wide-spacing arrays to narrow the null width in the direction of interference. Meanwhile, it avoids the physical implementation difficulties and mutual coupling effects associated with small-spacing array designs caused by the ultra-wideband characteristics of target signals. The spatio-temporal adaptive whitening filtering is then performed on each coprime-spacing subarray separately, adaptively forming two-dimensional nulls within the interference support region, thereby suppressing directional anthropogenic interference in analytical data before index calculation. Next, a frequency-dependent threshold scheme is utilized to obtain the binary spectrogram for each coprime-spacing subarray output, abating the influence from gain differences along the frequency axis for a certain direction. Afterwards, by leveraging the high spatial resolution of wide-spacing arrays and the interleaved characteristics of spatial aliasing null positions between the spatio-temporal frequency responses of the two subarrays with coprime spacing, a pointwise maximum fusion is applied to the above two binary spectrograms. This process reconstructs the binary time-frequency distribution structure of target signals outside the interference support region, leading to a single binary spectrogram where biological sound components are preserved to a great extent and anthropogenic interference is considerably suppressed. Ultimately, from this single binary spectrogram, the proportions of non-zero time-frequency bins within each frequency band are calculated and forwarded to the entropy function, resulting in the final AIC-FADI result. Results and Discussions The simulation result indicates that the proposed AIC-FADI maintains numerical robustness across an SINR range down to –15 dB (the yellow line in Fig. 5), substantially outperforming the classical ADI version based on single-channel noise reduction algorithm (FADI) and other ADI versions based on single-array interference suppression processing mentioned in this paper (AIC-FADI-s, AIC-FADI1, and AIC-FADI2). The real-world experiment confirms that the proposed spatio-temporal adaptive whitening filtering effectively suppresses wideband interference signals in complex scenarios, thereby improving the SINR of the analyzed recording. This enables some weaker biotic signals to exceed their corresponding frequency-dependent adaptive thresholds, greatly reducing missed detection of the target signal. In addition, by performing pointwise maximum fusion of the binary spectrograms from the two coprime-spacing subarray outputs, AIC-FADI further alleviates the extent of target signal missed detection (Fig. 8). Nevertheless, the real-world experiments also reveal that the interference suppression performance of AIC-FADI degrades for highly time-varying interference components. Conclusions This paper addresses the challenge of calculating acoustic indices reliably in complex soundscapes where directional anthropogenic interference overlaps with biotic signals in both time and frequency domains. A noise reduction strategy using coprime-spacing subarrays is proposed, and a new noise-robust acoustic index (AIC-FADI) is then developed. The method is evaluated through simulations and real-world recordings, and the results show that: (1) By applying spatio-temporal adaptive whitening filtering on each coprime-spacing subarray followed by pointwise maximum fusion, the proposed method achieves both wideband interference suppression capability and target information fidelity in complex soundscapes containing strong interference. (2) As a result, the proposed AIC-FADI maintains numerical robustness down to –15 dB SINR, substantially outperforming the classical FADI algorithm and other ADI versions based on single-array interference suppression methods. (3) The proposed method provides a feasible technical solution for extending the practical application scenarios and spatio-temporal coverage of biodiversity acoustic indices in human-dominated areas. However, this study only considers directional interference that is relatively stable or slowly time-varying. Hence, the interference suppression performance degrades for highly time-varying or uncorrelated noise components. These challenges should be addressed in future work through more advanced signal processing techniques to further improve the robustness of acoustic indices in highly complex acoustic environments.
- Microphone array /
- Adaptive interference cancellation /
- Spatio-temporal adaptive filtering /
- Spatial aliasing effect mitigation /
- Acoustic index

HTML全文

图 1 双元阵列空时自适应白化滤波流程示意图

下载: 全尺寸图片幻灯片

图 2 无混叠阵元间距下，双元阵列空时自适应白化滤波归一化空频幅度响应(干扰来向150°)

下载: 全尺寸图片幻灯片

图 3 宽间距条件下，双元阵列空时自适应白化滤波归一化空频幅度响应(干扰来向150°)

下载: 全尺寸图片幻灯片

图 4 自适应干扰对消依频声学多样性指数处理过程示意图

下载: 全尺寸图片幻灯片

图 5 五种ADI改进指数数值随SINR变化的仿真结果比较

下载: 全尺寸图片幻灯片

图 6 实测实验数据采集装置布局图

下载: 全尺寸图片幻灯片

图 7 参考阵元Mic1的实采信号时域波形图与时频谱图

下载: 全尺寸图片幻灯片

图 8 四种ADI改进指数的实采数据二值化时频谱图比较

下载: 全尺寸图片幻灯片

参考文献(16)

[1]	SONG Ziqi, YANG Zhichao, XIONG Yao, et al. Assessing the drivers of bird diversity in urban parks during winter: Insights from acoustic indices[J]. Ecological Indicators, 2025, 178: 113854. doi: 10.1016/j.ecolind.2025.113854.
[2]	XU Zhiyong, CHEN Lei, PIJANOWSKI B C, et al. A frequency-dependent acoustic diversity index: A revision to a classic acoustic index for soundscape ecological research[J]. Ecological Indicators, 2023, 155: 110940. doi: 10.1016/j.ecolind.2023.110940.
[3]	MAMMIDES C, PAN Wuyuan, HUANG Guohualing, et al. The combined effectiveness of acoustic indices in measuring bird species richness in biodiverse sites in Cyprus, China, and Australia[J]. Ecological Indicators, 2025, 170: 113105. doi: 10.1016/j.ecolind.2025.113105.
[4]	PAN Wuyuan, GOODALE E, JIANG Aiwu, et al. The effect of latitude on the efficacy of acoustic indices to predict biodiversity: A meta-analysis[J]. Ecological Indicators, 2024, 159: 111747. doi: 10.1016/j.ecolind.2024.111747.
[5]	FAIRBRASS A J, RENNERT P, WILLIAMS C, et al. Biases of acoustic indices measuring biodiversity in urban areas[J]. Ecological Indicators, 2017, 83: 169–177. doi: 10.1016/j.ecolind.2017.07.064.
[6]	LAI Xiaotian, XU Zhiyong, CHEN Lei, et al. Noise impact suppression for acoustic complexity index[C]. The 4th International Conference on Electronic Information Engineering and Computer, Shenzhen, China, 2024: 32–36. doi: 10.1109/EIECT64462.2024.10867106.
[7]	陈蕾, 许志勇, 苏菩坤, 等. 依频声学多样性指数用于人类活动区域的适用能力[J]. 生物多样性, 2024, 32(10): 24286. doi: 10.17520/biods.2024286. CHEN Lei, XU Zhiyong, SU Pukun, et al. Exploring the application of frequency-dependent acoustic diversity index in human-dominated areas[J]. Biodiversity Science, 2024, 32(10): 24286. doi: 10.17520/biods.2024286.
[8]	OKAMOTO R and OGUMA H. ChirpArray: A low-cost, easy-to-construct microphone array for long-term ecoacoustic monitoring[J]. Methods in Ecology and Evolution, 2025, 16(2): 302–308. doi: 10.1111/2041-210X.14474.
[9]	HEATH B E, SUZUKI R, LE PENRU N P, et al. Spatial ecosystem monitoring with a multichannel acoustic autonomous recording unit (MAARU)[J]. Methods in Ecology and Evolution, 2024, 15(9): 1568–1579. doi: 10.1111/2041-210X.14390.
[10]	VILLANUEVA-RIVERA L J, PIJANOWSKI B C, DOUCETTE J, et al. A primer of acoustic analysis for landscape ecologists[J]. Landscape Ecology, 2011, 26(9): 1233–1246. doi: 10.1007/s10980-011-9636-9.
[11]	PIERETTI N, FARINA A, and MORRI D. A new methodology to infer the singing activity of an avian community: The acoustic complexity index (ACI)[J]. Ecological Indicators, 2011, 11(3): 868–873. doi: 10.1016/j.ecolind.2010.11.005.
[12]	DMOCHOWSKI J, BENESTY J, and AFFES S. On spatial aliasing in microphone arrays[J]. IEEE Transactions on Signal Processing, 2009, 57(4): 1383–1395. doi: 10.1109/TSP.2008.2010596.
[13]	何劲, 唐莽, 舒汀, 等. 阵元位置互质的线性阵列: 互耦分析和角度估计[J]. 电子与信息学报, 2022, 44(8): 2852–2858. doi: 10.11999/JEIT210489. HE Jin, TANG Mang, SHU Ting, et al. Linear coprime sensor location arrays: Mutual coupling effect and angle estimation[J]. Journal of Electronics & Information Technology, 2022, 44(8): 2852–2858. doi: 10.11999/JEIT210489.
[14]	DU Yuxi, CUI Weijia, BA Bin, et al. Adaptive beamforming algorithm for subarrays based on coprime array[C]. The 3rd International Conference on Electronics and Information Technology, Chengdu, China, 2024: 872–876. doi: 10.1109/EIT63098.2024.10762283.
[15]	GRAF S, HERBIG T, BUCK M, et al. Features for voice activity detection: A comparative analysis[J]. EURASIP Journal on Advances in Signal Processing, 2015, 2015: 91. doi: 10.1186/s13634-015-0277-z.
[16]	YANG Jian, TU Yuwei, LU Jian, et al. Robust adaptive beamforming based on subspace decomposition, steering vector estimation and correction[J]. IEEE Sensors Journal, 2022, 22(12): 12260–12268. doi: 10.1109/JSEN.2022.3174848.