Detecting defects, which is a branch of target detection in the field of computer vision, is widely used in factory production. To solve the problems in existing detection algorithms that relate to their insensitivity to large or medium defect targets on bearing covers, their difficulty in detecting subtle defects effectively and their lack of real-time detection, in this work, we establish a large-scale bearing-cover defect dataset and propose an improved YOLOv3 network model. The proposed model is divided into four submodels: the bottleneck attention network (BNA-Net), the attention prediction subnet model, the defect localization subnet model, and the large-size output feature branch. To test the generality, robustness and practicability of the new model, we design a comparative experiment under abnormal illumination conditions. We design an ablation experiment to verify the validity of the proposed submodules. The experimental results show that our model solves the problem of the YOLOv3 algorithm’s insensitivity to medium or large targets and satisfies real-time detection conditions. The mAP result is 69.74%, which is 16.31%, 13.4%, 13%, 10.9%, and 7.2% more than that of YOLOv3, EfficientDet-D2, YOLOv5, YOLOv4, and PP-YOLO, respectively.

Few-shot object detection (FSOD) addresses the challenge of object recognition under limited annotation conditions, offering practical advantages for smart agriculture, where large-scale labeling of diverse fruit cultivars is often infeasible. To handle the visual complexity of orchard environments—such as occlusion, subtle morphological differences, and dense foliage—this study presents a lightweight tri-modal fusion framework. The model initially employs a CLIP-based semantic prompt encoder to extract category-aware cues, which guide the Segment Anything Model (SAM) in producing structure-preserving masks. These masks are then incorporated via a Semantic Fusion Module (SFM): a Mask-Saliency Adapter (MSA) and a Feature Enhancement Recomposer (FER), enabling spatially aligned and semantically enriched feature modulation. An Attention-Aware Weight Estimator (AWE) further optimizes the fusion by adaptively balancing semantic and visual streams using global saliency cues. The final predictions are subsequently generated by a YOLOv12 detection head. Experiments conducted on four fruit detection benchmarks—Cantaloupe.v2, Peach.v3, Watermelon.v2, and Orange.v8—demonstrate that the proposed method consistently surpasses five representative FSOD baselines. Performance improvements include +7.9% AP@0.5 on Cantaloupe.v2, +5.4% Precision on Peach.v3, +7.4% Precision on Watermelon.v2, and +5.9% AP@0.75 on Orange.v8. These results underscore the model’s effectiveness in orchard-specific scenarios and its potential to facilitate cultivar identification, digital recordkeeping, and cost-efficient agricultural monitoring.

Mechanical fault can cause economic loss of different degrees, even casualties. Timely fault diagnosis is an essential condition for ensuring safe production in modern industry. With the growth of intelligent manufacturing, more and more attention is...

In response to the demand for efficient and accurate recognition of traffic police gestures by driverless vehicles, this paper introduces a novel traffic police gesture recognition framework (Novel Traffic Police Gesture Recognizer, NTPGR). Initially, keypoints related to traffic police gestures are extracted using the Efficient Progressive Feature Fusion Network (EPFFNet), followed by feature modeling and fusion to enable the recognition network to better learn the temporal characteristics of gestures. Additionally, a convolution network branch and a hybrid attention branch are incorporated to further extract skeleton information from the traffic police gesture data, assign different temporal weights to key frames, and enhance the focus on important channels. Finally, in conjunction with Long Short Term Memory (LSTM), a multi-branch gesture recognition network, termed the Multi-Sequence Gesture Recognition Network (MSNet), is proposed to facilitate the integration of three branches of gesture features, thereby enhancing the targeted extraction of temporal characteristics in traffic police gestures. Experimental results indicate that NTPGR achieves 97.56% and 96.76% accuracy on the Police Gesture Dataset and UTD-MHAD Dataset, respectively, as well as average response times of 0.76s and 0.74s. It not only recognizes traffic police gestures in real-time with high efficiency but also demonstrates strong robustness and Credibility in recognizing gestures in complex environments and dynamic scenarios.