Robust single-pixel imaging for high-speed rotating objects with large rotational speed variations

Long-term imaging for high-speed rotating objects holds promising application prospects for video monitoring of rotating mechanical equipment. Single-pixel imaging exhibits distinct advantages in terms of data transmission bandwidth and storage-capacity requirements. Nevertheless, single-pixel imaging of high-speed rotating objects under large rotational speed variations still faces challenges. Herein, we propose a robust single-pixel imaging method suitable for high-speed rotating objects with drastic rotational speed fluctuations. During data acquisition, positioning signals from a rotation angle positioning module are used to synchronously trigger the projector for pattern projection. In the image reconstruction procedure, an adaptive algorithm is utilized to align the sampling values acquired by the single-pixel detector according to rotation angles. Experiments demonstrate single-pixel imaging of a high-speed rotating fan with rotational speeds varying from 6485.1 to 15205.3 revolutions per minute. A total of 7892 images were reconstructed at different rotation angles over one full revolution of the fan. Furthermore, by utilizing the proposed method with an undersampling approach, we experimentally achieved robust imaging of the cooling fan. Experimental results verify that the proposed method is robust against large rotational speed variations, and it is expected to find applications such as online long-term video monitoring of rotating machinery.