In order to deal with the shortcomings (such as poor robustness) of the traditional single-channel vibration signal in the comprehensive monitoring of the gearbox fault state, a multi-channel decision-level fusion algorithm was proposed based on symmetrized dot pattern (SDP) analysis, with the visual geometry group 16 network (VGG16) fault diagnosis model. Firstly, the SDP method was used to convert the vibration signal of a single multi-channel sensor into an imaging arm. Secondly, the obtained image arm was input into the VGG16 convolutional neural network in order to train the fault diagnosis model that can be obtained. Then, the SDP images of the signals that were to be measured from multiple multi-channel sensors were input into the fault diagnosis model, and the diagnosis results of multiple multi-channel sensors could then be obtained. Experimentally, it was demonstrated that the diagnostic results of multi-channel sensors one, two, and three were more accurate than those of single-channel sensors one, two, and three, by 3.01%, 16.7%, and 5.17%, respectively. However, the fault generation was not generated in a single direction, but rather multiple directions. In order to improve the comprehensiveness of the raw vibration data, a fusion method using DS (Dempster–Shafer) evidence theory was proposed in order to fuse multiple multi-channel sensors, in which the accuracy achieved 99.93% when sensor one and sensor two were fused, which was an improvement of 8.88% and 1.02% over single sensors one and two, respectively. When sensor one and sensor three were fused, the accuracy reached 99.31%, which was an improvement of 8.31% and 6.17% over single sensors one and three, respectively. When sensor two and sensor three were fused, the accuracy reached 99.91%, which was an improvement of 1.00% and 6.74% over single sensors two and three, respectively. When three sensors were fused simultaneously, the accuracy reached 99.99%, which was 8.93%, 1.08%, and 6.81% better than single sensors one, two, and three, respectively. Therefore, it can be proved that the number of sensor channels has a great influence on the diagnosis results.

A decision-level approach using multi-sensor-based symmetry dot pattern (SDP) analysis with a Visual Geometry Group 16 network (VGG16) fault diagnosis model for multi-source information fusion was proposed to realize accurate and comprehensive fault diagnosis of gearbox gear teeth. Firstly, the SDP technique was used to perform a feature-level fusion of the fault states of gearbox gear collected by multiple sensors, which could initially visualize the vibration states of the gear teeth in different states. Secondly, the SDP images obtained were combined with the deep learning VGG16. In this way, the local diagnostic results of each sensor can be easily obtained. Finally, the local diagnostic results of each sensor were combined with the DS evidence theory to achieve decision-level fusion, which can better realize comprehensive fault detection for gearbox gear teeth. Before fusion, the accuracies of the three sensors were 96.43%, 93.97%, and 93.28%, respectively. When sensor 1 and sensor 2 were fused, the accuracy reached 99.93%, which is 3.52% and 6.34% better than when using sensors 1 and 2, respectively, alone. When sensor 1 and sensor 3 were fused, the accuracy reached 99.96%, marking an improvement of 3.36% and 6.85% over individual use of sensors 1 and 3, respectively. When sensor 2 and sensor 3 were fused, the accuracy reached 99.40%, which is 5.78% and 6.56% better than individual use of sensors 2 and 3, respectively. When the three sensors were fused simultaneously, the accuracy reached 99.98%, which is 3.68%, 6.40%, and 7.18% better than individual use of sensors 1, 2, and 3, respectively.

In order to accurately realize the contact fatigue state identification of specimen, a new method based on vibration and image heterogeneous data, as well as on D-S evidence theory, is proposed. Firstly, combined with the bearing public data set from CWRU, the vibration signal imaging methods such as SDP, GAF and GRI, as well as neural network models such as VGG16, ResNet and S-T, were compared and analyzed. It is determined that the SDP method is used to visualize the vibration signal, and the two state identification evidence bodies based on the vibration information source are obtained through the VGG16 and ResNet models. Secondly, combined with image monitoring signals, the fatigue defect identification method based on automatic weighting threshold and the detection error dynamic compensation method based on fatigue defect edge features are used to quantify the fatigue damage area and obtain the state identification evidence body based on the image information source. On this basis, a state identification network model based on vibration and image spatiotemporal heterogeneous data is constructed, and the D-S evidence theory is used to realize the contact fatigue state identification of the specimen. The results show that fusion of vibration and image data can achieve information complementarity and may identify the contact fatigue state of the specimen more accurately. The accuracy of state identification after fusion is 98.67%, which is at least 3% higher than that of a single information source. This research is of great significance for the accurate acquisition of material contact fatigue properties and has certain reference value for the heterogeneous data fusion from different sources.