Accurate detection of wood defects plays a crucial role in optimizing wood utilization, minimizing corporate expenses, and safeguarding precious forest resources. To achieve precise identification of surface defects in wood, we present a novel approach called the Omni-dynamic convolution coordinate attention-based YOLO (ODCA-YOLO) model. This model incorporates an Omni-dimensional dynamic convolution-based coordinate attention (ODCA) mechanism, which significantly enhances its ability to detect small target defects and boosts its expressiveness. Furthermore, to reinforce the feature extraction and fusion capabilities of the ODCA-YOLO network, we introduce a highly efficient features extraction network block known as S-HorBlock. By integrating HorBlock into the ShuffleNet network, this design optimizes the overall performance. Our proposed ODCA-YOLO model was rigorously evaluated using an optimized wood surface defect dataset through ablation and comparison experiments. The results demonstrate the effectiveness of our approach, achieving an impressive 78.5% in the mean average precision (mAP) metric and showing a remarkable 9% improvement in mAP compared to the original algorithm. Our proposed model can satisfy the need for accurate detection of wood surface defects.

The detection of wood defect is a crucial step in wood processing and manufacturing, determining the quality and reliability of wood products. To achieve accurate wood defect detection, a novel method named BPN-YOLO is proposed. The ordinary convolution in the ELAN module of the YOLOv7 backbone network is replaced with Pconv partial convolution, resulting in the P-ELAN module. Wood defect detection performance is improved by this modification while unnecessary redundant computations and memory accesses are reduced. Additionally, the Biformer attention mechanism is introduced to achieve more flexible computation allocation and content awareness. The IOU loss function is replaced with the NWD loss function, addressing the sensitivity of the IOU loss function to small defect location fluctuations. The BPN-YOLO model has been rigorously evaluated using an optimized wood defect dataset, and ablation and comparison experiments have been performed. The experimental results show that the mean average precision (mAP) of BPN-YOLO is improved by 7.4% relative to the original algorithm, which can better meet the need to accurately detecting surface defects on wood.

Detection and reconstruction of actuator faults in a flight control system are crucial to avoid negative impacts on the aircraft itself, as well as human and environmental systems. In this study, an actuator fault detection and reconstruction scheme based on fault classification for Hex-rotor unmanned aerial vehicle is designed. First, the fault type of actuator is analyzed and classified, then an actuator fault model is established based on multiple fault classification. Secondly, a fault detection and reconstruction (FDR) scheme for the Hex-rotor unmanned aerial vehicle is proposed. In the proposed scheme, a fault observer group based on an extended Kalman Filter (EKF) is designed for fault detection and isolation, and the flight state feedback required by the fault observer group is provided by the multi-sensor navigation unit (MSNU). Then a fault reconstruction algorithm of actuator is proposed by using the output of the fault observer group. The designed FDR system is applied to Hex-rotor unmanned aerial vehicle prototype for its evaluation. The simulation results show that the proposed design is capable of fault detection and the reconstruction of actuator faults, and the actual flight verifies the effectiveness of the proposed methods.

The detection of defects on the surface is of great importance for both the production and the application of strip steel. In order to detect the defects accurately, an improved YOLOv7-based model for detecting strip steel surface defects is developed. To enhances the ability of the model to extract features and identify small features, the ConvNeXt module is introduced to the backbone network structure, and the attention mechanism is embedded in the pooling module. To reduce the size and improves the inference speed of the model, an improved C3 module was used to replace the ELAN module in the head. The experimental results show that, compared with the original models, the mAP of the proposed model reached 82.9% and improved by 6.6%. The proposed model can satisfy the need for accurate detection and identification of strip steel surface defects.

Strong background noise and complicated interfering signatures when implementing vibration-based monitoring make it difficult to extract the weak diagnostic features due to incipient faults in a multistage gearbox. This can be more challenging when multiple faults coexist. This paper proposes an effective approach to extract multi-fault features of a wind turbine gearbox based on an integration of minimum entropy deconvolution (MED) and multipoint optimal minimum entropy deconvolution adjusted (MOMEDA). By using simulated periodic transient signals with different noise to signal ratios (SNR), it evaluates the outstanding performance of MED in noise suppression and reveals the deficient in extract multiple impulses. On the other hand, MOMEDA can performs better in extracting multiple pulses but not robust to noise influences. To compromise the merits of them, therefore the diagnostic approach is formalized by extracting the multiple weak features with MOMEDA based on the MED denoised signals. Experimental verification based on vibrations from a wind turbine gearbox test bed shows that the approach allows successful identification of multiple faults occurring simultaneously on the shaft and bearing in the high speed transmission stage of the gearbox.

Wood surface defect detection is a challenging task due to the complexity and variability of defect types. To address these challenges, this paper introduces a novel deep learning approach named SiM-YOLO, which is built upon the YOLOv8 object detection framework. A fine-grained convolutional structure, SPD-Conv, is introduced with the aim of preserving detailed defect information during the feature extraction process, thus enabling the model to capture the subtle variations and complex details of wood surface defects. In the feature fusion stage, a SiAFF-PANet-based wood defect feature fusion module is designed to improve the model’s ability to focus on local contextual information and enhance defect localization. For classification and regression tasks, the multi-attention detection head (MADH) is employed to capture cross-channel information and the accurate spatial localization of defects. In addition, MPDIoU is employed to optimize the loss function of the model to reduce the leakage of detection due to defect overlap. The experimental results show that SiM-YOLO achieves superior performance compared to the state-of-the-art YOLO algorithm, with a 9.3% improvement in mAP over YOLOX and a 4.3% improvement in mAP over YOLOv8. The Grad-CAM visualization further illustrates that SiM-YOLO provides more accurate defect localization and effectively reduces misdetection and omission issues. This study highlights the effectiveness of SiM-YOLO for wood surface defect detection and offers valuable insights for future research and practical applications in quality control.