Objects classification generally relies on image acquisition and analysis. Real-time classification of high-speed moving objects is challenging, as both high temporal resolution in image acquisition and low computational complexity in objects classification algorithms are required. Here we propose and experimentally demonstrate an approach for real-time moving objects classification without image acquisition. As objects classification algorithms rely on the feature information of objects, we propose to use spatial light modulation to acquire the feature information directly rather than performing image acquisition followed by features extraction. A convolutional neural network is designed and trained to learn the spatial features of the target objects. The trained network can generate structured patterns for spatial light modulation. Using the resulting structured patterns for spatial light modulation, the feature information of target objects can be compressively encoded into a short light intensity sequence. The resulting one-dimensional signal is collected by a single-pixel detector and fed to the convolutional neural network for objects classification. As experimentally demonstrated, the proposed approach can achieve accurate and real-time classification of fast moving objects. The proposed method has potential applications in the fields where fast moving objects classification in real time and for long duration is required.

Due to limited data transmission bandwidth and data storage space, it is challenging to perform fast-moving objects classification based on high-speed photography for a long duration. Here we propose a single-pixel classification method with deep learning for fast-moving objects. The scene image is modulated by orthogonal transform basis patterns, and the modulated light signal is detected by a single-pixel detector. Thanks to the property that the natural images are sparse in the orthogonal transform domain, we used a small number of basis patterns of discrete-sine-transform to obtain feature information for classification. The proposed neural network is designed to use single-pixel measurements as network input and trained by simulation single-pixel measurements based on the physics of the measuring scheme. Differential measuring can reduce the difference between simulation data and experiment data interfered by slowly varying noise. In order to improve the reliability of the classification results for fast-moving objects, we employed a measurement data rolling utilization approach for repeated classification. Long-duration classification of fast-moving handwritten digits that pass through the field of view successively is experimentally demonstrated, showing that the proposed method is superior to human vision in fast-moving digit classification. Our method enables a new way for fast-moving object classification and is expected to be widely implemented.