In response to challenges associated with feature extraction and diagnostic models’ complexity in the early diagnosis of bearings’ faults, this paper presents an innovative approach for the early fault diagnosis of rolling bearings. This method combined concepts from frequency domain signal analysis with lightweight neural networks. To begin, vibration signals from rolling bearings were collected using vibration sensors, and the mean square value was utilized as an indicator for accurate early fault signal extraction. Subsequently, employing the fractional Fourier transform, the time domain signal was converted into a frequency domain signal, which provided more detailed frequency feature information. The fusion process combined amplitude frequency and phase frequency information, and was visualized as a Gram angle field map. The lightweight neural network Xception was selected as the primary fault diagnosis tool. Xception, a convolutional neural network (CNN) variant, was chosen for its lightweight design, which maintains excellent performance while significantly reducing model parameters. The experimental results demonstrated that the Xception model excelled in rolling bearing fault diagnosis, particularly when utilizing fused information datasets. This outcome underscores the advantages of combining information fusion and the Xception model to enhance the accuracy of early rolling bearing fault diagnosis, and offers a viable solution for health monitoring and fault diagnosis in industrial settings.

Rolling bearings are prone to failure, meaning that research on intelligent fault diagnosis is crucial in relation to this key transmission component in rotating machinery. The application of deep learning (DL) has significantly advanced the development of intelligent fault diagnosis. This paper proposes a novel method for rolling bearing fault diagnosis based on the fractional constant Q non-stationary Gabor transform (FCO-NSGT) and VMamba-Conv. Firstly, a rolling bearing fault experimental platform is established and the vibration signals of rolling bearings under various working conditions are collected using an acceleration sensor. Secondly, a kurtosis-to-entropy ratio (KER) method and the rotational kernel function of the fractional Fourier transform (FRFT) are proposed and applied to the original CO-NSGT to overcome the limitations of the original CO-NSGT, such as the unsatisfactory time–frequency representation due to manual parameter setting and the energy dispersion problem of frequency-modulated signals that vary with time. A lightweight fault diagnosis model, VMamba-Conv, is proposed, which is a restructured version of VMamba. It integrates an efficient selective scanning mechanism, a state space model, and a convolutional network based on SimAX into a dual-branch architecture and uses inverted residual blocks to achieve a lightweight design while maintaining strong feature extraction capabilities. Finally, the time–frequency graph is inputted into VMamba-Conv to diagnose rolling bearing faults. This approach reduces the number of parameters, as well as the computational complexity, while ensuring high accuracy and excellent noise resistance. The results show that the proposed method has excellent fault diagnosis capabilities, with an average accuracy of 99.81%. By comparing the Adjusted Rand Index, Normalized Mutual Information, F1 Score, and accuracy, it is concluded that the proposed method outperforms other comparison methods, demonstrating its effectiveness and superiority.