KEYWORDS: Fluctuations and noise, Data modeling, Image enhancement, Defect detection, Image processing, Image classification, Detection and tracking algorithms, Neurons, Feature extraction, Education and training
At present, the traditional recognition methods such as naked eye recognition and machine vision are mainly used for the surface defect image recognition of fan blades, and their robustness is poor. In order to improve the recognition accuracy of fan blade surface defects, an improved ResNet50 algorithm is proposed in this paper. 1416 fan blade images were captured by UAV, and the data set of fan blade surface defects was established through screening, classification and data expansion. In order to verify the validity of the experiment, the data sets of sand holes and oil pollution defects of fan blades were subsequently supplemented in this experiment, and the same data set processing method was used. With ResNet50 as the backbone feature extraction network, the channel attention mechanism SENet is introduced to enhance the reuse rate of useful features and suppress useless features. In order to reduce the risk of inactivation of some neurons using ReLu activation function, this paper uses Leaky-ReLU activation function to redesign the residual block. The experimental results show that the classification accuracy of the algorithm on the self-made data is up to the baseline model, which can meet the actual demand.
To solve the problem of adhesion objects and data distribution deviations in few-shot scenarios, a synthetic aperture radar (SAR) object detection method based on meta-learning is proposed, which includes support feature guidance block and variational inference block. The former enhances the key features used for bounding box positioning in the query feature, so that the module can generate accurate proposals even in face of the adherent SAR objects. On this basis, to correct the deviation of the data distribution caused by the few-shot data, a variational inference block is constructed to map the supporting features to the class distribution in the hidden space. To fuse robust class-level features, meta-knowledge is used to calculate the distribution of the support feature classes of classes. The proposed algorithm uses a few-shot support set data to migrate priori knowledge to a class using the few-shot tasks and data double sampling. Moreover, a few-shot SAR object detection dataset is established to verify the effectiveness of the proposed method, and the experimental results show that our method has obvious advantages over the representative few-shot SAR object detection algorithms.
With the development of synthetic aperture radar (SAR) technology, more SAR datasets with high resolution and large scale have been obtained. Research using SAR images to detect and monitor marine targets has become one of the most important marine applications. In recent years, deep learning has been widely applied to target detection. However, it was difficult to use deep learning to train an SAR ship detection model in complex scenes. To resolve this problem, an SAR ship detection method combining YOLOv4 and the receptive field block (CY-RFB) was proposed in this paper. Extensive experimental results on the SAR-Ship-Dataset and SSDD datasets demonstrated that the proposed method had achieved supreme detection performance compared to the state-of-the-art ship detection methods in complex scenes, whether they were in offshore or inshore scenes of SAR images.
Motion trajectory recognition is one of the most important means to determine the identity of a moving object. A compact and discriminative feature representation method can improve the trajectory recognition accuracy. This paper presents an efficient framework for action recognition using a three-dimensional skeleton kinematic joint model. First, we put forward a rotation-scale-translation-invariant shape descriptor based on point context (PC) and the normal vector of hypersurface to jointly characterize local motion and shape information. Meanwhile, an algorithm for extracting the key trajectory based on the confidence coefficient is proposed to reduce the randomness and computational complexity. Second, to decrease the eigenvalue decomposition time complexity, a tensor shape descriptor (TSD) based on PC that can globally capture the spatial layout and temporal order to preserve the spatial information of each frame is proposed. Then, a multilinear projection process is achieved by tensor dynamic time warping to map the TSD to a low-dimensional tensor subspace of the same size. Experimental results show that the proposed shape descriptor is effective and feasible, and the proposed approach obtains considerable performance improvement over the state-of-the-art approaches with respect to accuracy on a public action dataset.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.