School of Computer Science and Engineering, Xi'an University of Technology, Xi'an 710048, China.
Shaanxi Key Laboratory for Network Computing and Security Technology, Xi'an 710048, China.
Sensors (Basel). 2018 Nov 15;18(11):3981. doi: 10.3390/s18113981.
Crowd counting is of significant importance for numerous applications, e.g., urban security, intelligent surveillance and crowd management. Existing crowd counting methods typically require specialized hardware deployment and strict operating conditions, thereby hindering their widespread application. To acquire a more effective crowd counting approach, a device-free counting method based on Channel Status Information (CSI) is proposed. The wavelet domain denoising is introduced to mitigate environment noise. Furthermore, the amplitude or phase covariance matrix is extracted as the eigenmatrix. Moreover, both the spatial diversity and frequency diversity are leveraged to improve detection robustness. At the same experimental environment, the accuracy of the proposed CSI-based method is compared with a renowned crowd counting one, i.e., Electronic Frog Eye: Counting Crowd Using WiFi (FCC). The experimental results reveal an accuracy improvement of 30% over FCC.
人群计数对于许多应用都非常重要，例如城市安全、智能监控和人群管理。现有的人群计数方法通常需要专门的硬件部署和严格的操作条件，从而阻碍了它们的广泛应用。为了获得更有效的人群计数方法，提出了一种基于信道状态信息（CSI）的无设备计数方法。引入小波域去噪来减轻环境噪声。此外，提取幅度或相位协方差矩阵作为特征矩阵。此外，还利用空间多样性和频率多样性来提高检测鲁棒性。在相同的实验环境下，将所提出的基于 CSI 的方法与著名的人群计数方法（即基于 WiFi 的电子青蛙眼：计数人群（FCC））进行了准确性比较。实验结果表明，该方法的准确性比 FCC 提高了 30%。
