A study on vehicle bottom safety detection technology based on image feature matching algorithm

Objective: We propose a deep-learning-based underwater target detection system that can effectively solve the problem of underwater optical image target detection and recognition. Methods: In this paper, based on the depth of the underwater optical image target detection and recognition and using a learning model, we put forward corresponding solutions using the concept of style migration solutions, such as training samples. A lack of variability and poor generalization of practical applications presents a challenge for underwater object identification. The UW_YOLOv3 lightweight model was proposed to solve the problems of calculating energy consumption and storage resource limitations in underwater application scenarios. The detection and recognition module, based on deep learning, can deal with the degradation process of underwater imaging by embedding an image enhancement module into the detection and recognition module for the joint tuning and transferring of knowledge. Results: The detection accuracy of the UW_YOLOv3 model designed in this paper outperformed the lightweight algorithm YOLOV3-TINY by 7.9% at the same image scale input. Compared with other large algorithms, the detection accuracy was lower, but the detection speed was much higher. Compared with the SSD algorithm, the detection accuracy was only 4.7 lower; the speed was 40.9 FPS higher; and the rate was nearly 16 times higher than Faster R-CNN. When the input scale was 224, although part of the accuracy was lost, the detection speed doubled, reaching 156.9 FPS. Conclusion: Based on our framework, the problem of underwater optical image target detection and recognition can be effectively solved. Relevant studies have not only enriched the theory of target detection and glory, but have also provided optical glasses with a clear vision for appropriate underwater application systems.

Multifeature decision fusion is an effective way to promote the performance of target recognition of synthetic aperture radar (SAR) images. This paper proposes a joint multifeature decision fusion strategy for target recognition in SAR images based on multitask compressive sensing (MtCS). The proposed method can exploit the intercorrelations among different features by enforcing the constraint on the sparsity pattern. Furthermore, the time consumption for MtCS is almost the same with that of single feature-based compressive classification, such as sparse representation-based classification. Experiments on the moving and stationary target acquisition and recognition dataset and comparison with several state-of-the-art methods demonstrate the validity of the proposed method.

Image target detection and recognition is an important research area in image understanding and computer vision, and has attracted extensive attention of scholars for its wide application. In recent years, with the enlightenment of human vision system, deep learning technology has been put forward, and has become a hot spot. The deep learning model has a multi-layer network structure, which can overcome the shortcomings of the traditional image target recognition methods. In the detection and recognition of image targets, deep learning methods are more and more widely used. This is because the deep learning multilayer network structure can express complex functions in a more concise way than the shallow network, and can learn the deep feature representation. Therefore, this paper uses the method based on deep learning, researches on image target detection and recognition, mainly introduces the region proposal based on the depth of learning based target recognition method of support vector machine algorithm and target detection depth. In this paper, deep learning is applied to image detection and recognition, further to improve the accuracy of image detection and recognition. It has important guiding significance in deep learning and image research.

Synthetic aperture radar (SAR) image target detection and recognition (SAR-TDR) tasks have become research hot spots in the remote sensing application. These targets include ships, vehicles, aircraft, oil tanks, bridges, and so on. However, with the rapid development of SAR technology and increasingly complex electromagnetic environment, complex characteristics of SAR images bring severe challenges to the accurate SAR-TDR via traditional physical models or manual feature-extraction-based machine learning methods. In recent years, deep learning (DL), as a powerful automatic feature extraction algorithm, has been widely used in the computer vision domain. More specifically, DL has also been introduced into the SAR-TDR tasks and has effectively achieved good performance in terms of accuracy, real-time processing, etc. With the rapid development of DL, SAR image processing, and practical requirements of SAR-TDR in civilian and military domains, it is crucial to conduct a systematic survey on SAR-TDR in the past few years. In this survey article, we mainly conduct a systematic overview of DL-based SAR-TDR literature on two tasks, i.e., target recognition (e.g., ground vehicles, ships, and aircraft) and target detection (e.g., ships, aircraft, change detection, sea surface oil spills, and oil tanks). More specifically, our related works about these topics are also presented to verify the effectiveness of DL-based methods. First, several DL methods (e.g., convolutional neural networks, recurrent neural networks, autoencoders, and generative adversarial networks), commonly used in SAR-TDR, are briefly introduced. Then, a systematic review of DL-based SAR-TDR (including our related works) is presented. Finally, the current challenges and future possible research directions are deeply analyzed and discussed.

The mining belt conveyor is one of the most important modules in coal mine, whose safety always be threatened by the foreign objects. Although the traditional target detection methods achieve promising results in various computer vision tasks, the performance heavily depends on sufficient labelled data. However, in real-world production scenario, it is difficult to acquire huge number of images with foreign objects. The obtained datasets lacking of capacity and diversity are not suitable for training supervised learning-based foreign objects detection models. To address this concern, we propose a novel method for detecting the foreign objects on the surface of underground coal conveyor belt via improved GANomaly. The proposed foreign objects detection method employs generative adversarial networks (GAN) with attention gate to capture the distribution of normality in both high-dimensional image space and low-dimensional latent vector space. Only the normal images without foreign object are utilized to adversarially train the proposed network, including a U-shape generator to reconstruct the input image and a discriminator to classify real images from reconstructed ones. Then the combination of the difference between the input and generated images as well as the difference between latent representations are utilized as the anomaly score to evaluate whether the input image contain foreign objects. Experimental results over 707 images from real-world industrial scenarios demonstrate that the proposed method achieves an area under the receiver operating characteristic curve of 0.864 and is superior to the previous GAN-based anomaly detection methods.

Airport runway foreign object detection systems can quickly and accurately detect and identify foreign runway objects, which is significant for ensuring airport flight safety. Because of the drawbacks shown by the algorithm, the paper proposes to combine a new system scheme based on deep learning to obtain multiple feature information for identifying foreign objects on airport runways and improve the recognition accuracy of foreign object detection. This paper designs and constructs a dataset for accomplishing airport runway foreign object detection based on the data distribution and attribute semantics of actual airport runway foreign object scenarios and the technical features of deep learning, designs FOD detection and multi-attribute recognition networks, further design algorithms, and perform validation. The results show that the deep learning technology can accomplish all tasks of the airport runway foreign object detection system, which has not only good robustness to different environments but also has practical value for multi-tasking, and the localization task can accurately obtain the location information of foreign objects and improve the recognition accuracy of foreign object detection. Therefore, the deep learning-based airport runway foreign object recognition system designed in this paper is effective and can improve the accuracy of foreign object recognition.

In this study, three decision fusion strategies in synthetic aperture radar (SAR) image target recognition for classification of ground vehicles in the Moving and Stationary Target Acquisition and Recognition public release database are proposed. These strategies include multi-view decision fusion strategy, multi-feature decision fusion strategy and multi-classifier decision fusion strategy. In multi-view decision fusion strategy, each view implements its recognition process individually and provides only its decision from the classifier to the fuser. In multi-feature decision fusion strategy, a variety of feature extraction algorithms are used to extract features from an image separately and each of the extracted feature vectors is submitted to a classifier. The decisions from the classifiers are then provided to the fuser. In multi-classifier decision fusion strategy, different classifiers are used to classify a feature vector. Each classifier executes its classification process individually and then provides its decision to the fuser. The fuser combines decisions to produce a single fused final decision in these three strategies. The three proposed strategies are evaluated and verified with experiments. Experimental results indicate that there are significant target recognition performance benefits in the probability of correct classification, when the proposed decision fusion strategies are properly applied in SAR image target recognition.

Synthetic aperture radar (SAR) image target recognition technology is aimed at automatically determining the presence or absence of target information from the input SAR image and improving the efficiency and accuracy of SAR image interpretation. Based on big data analysis, dirty data is removed, clean data is returned, and standardized processing of SAR image data is realized. At the same time, by establishing a statistical model of coherent speckles, the convolutional autoencoder is used to denoise the SAR image. Finally, the network model modified by softmax cross‐entropy loss and Fisher loss is used for automatic target recognition. Based on the MSTAR data set, two scene graphs containing the target synthesized by the background image and the target slice are used for experiments. Several comparative experiments have verified the effectiveness of the classification and recognition model in this paper.

The goal of the present research is to despeckle SAR images, which is critical for segmentation and target recognition in satellite SAR images. When a despeckling algorithm is applied to a SAR image, important information such as the edges, corners, textures, and object parts will degrade. Curvelet transform is a recently proposed form of multi-scale analysis that achieves better performance of wavelet and Gabor transforms in edge and curve detection. This is a geometric transform that is useful for SAR image processing. For unsupervised texture images, segmentation is different and distinct from the textures, so the textures at the boundary noises will disappear. Curvelet transform has produced good results in the detection of curved edges with higher accuracy in finding the orientation than wavelet transforms. The present study uses fast discrete curvelet transform (FDCT) based on wresting and uses unsupervised adaptive threshold learning to develop a new despeckling algorithm for SAR images. In the proposed algorithm, each segment of the SAR image can be learned for selection of its adaptive threshold. Simulation results demonstrate that the proposed algorithm performs better than similar methods.

Background:: The image processing technology can be adopted in the power transmission incident detection system, when it is combined with artificial intelligence and machine learning academic knowledge of power transmission surveillance video analysis system, it can automatically detect abnormal objects in the power transmission system. It can quickly and accurately detect abnormal objects, known as environmental object feature detection necessary for the safety of power transmission system. Objective:: In order to improve the object detection ability of the power transmission system, it adopts artificial intelligence and machine learning knowledge in the power transmission surveillance system and it can automatically detect abnormal events or objects. Methods:: Compared with differential binary target detection technique, it proposes a new adaptive background subtraction threshold algorithm to adapt the complex power transmission monitoring environment. Meanwhile, it takes special recognition algorithm for power transmission monitoring environment to ensure the accuracy and stability of the detection system. Results:: The proposed method can be used in the monitoring system and the accuracy and stability of the detection system have been verified through experiments. Conclusion:: Through adopting the special recognition algorithm for power transmission monitoring environment, it can ensure the accuracy and stability of the detection system, in the detection experiments, it can analyze the objects more easily than before.

Synthetic Aperture Radar (SAR) images are prone to be contaminated by noise, which makes it very difficult to perform target recognition in SAR images. Inspired by great success of very deep convolutional neural networks (CNNs), this paper proposes a robust feature extraction method for SAR image target classification by adaptively fusing effective features from different CNN layers. First, YOLOv4 network is fine-tuned to detect the targets from the respective MF SAR target images. Second, a very deep CNN is trained from scratch on the moving and stationary target acquisition and recognition (MSTAR) database by using small filters throughout the whole net to reduce the speckle noise. Besides, using small-size convolution filters decreases the number of parameters in each layer and, therefore, reduces computation cost as the CNN goes deeper. The resulting CNN model is capable of extracting very deep features from the target images without performing any noise filtering or pre-processing techniques. Third, our approach proposes to use the multi-canonical correlation analysis (MCCA) to adaptively learn CNN features from different layers such that the resulting representations are highly linearly correlated and therefore can achieve better classification accuracy even if a simple linear support vector machine is used. Experimental results on the MSTAR dataset demonstrate that the proposed method outperforms the state-of-the-art methods.

Target recognition in thermal infrared images is challenging due to high variability of target IR signature and competing background IR signature due to a number of environmental and target parameters. Traditional hand-crafted feature extractors are limited by these challenges. Recently, deep learning has shown promising success for a number of computer vision works. In this paper, deep CNN-based feature extraction is explored for target recognition in thermal images. In this study, two pre-trained CNNs, AlexNet and VGG19 are considered. A number of deep CNN-based feature extractors are evaluated by extracting features from different layers of the network. The results indicate the robustness of the deep CNN-based feature extractor. The VGG19_fc6 architecture has demonstrated superior performance with 6% improvement in the classification accuracy against the WignerMSER based state of the art target recognition on two class FLIR thermal infrared dataset.

With the progress of human space technology, mankind stepped into more distant space. Due to the long distance from the earth in deep space exploration, higher requirements were put forward for intelligent cognition of targets and environment. In this paper, the space target classification and recognition technology based on deep learning was studied by taking the classification of three types of satellites as an example. A satellite simulation sample set for deep learning was established, and a ResNet multi-layer convolutional neural network model suitable for spatial target characteristics was constructed. The training and test of satellite intelligent classification were completed, and the feature extraction results of the neural network were visualized. The accuracy rate of satellite classification identification for the remote sensing satellites, communication satellites and navigation satellites reached 90%, which provided a reference for the development of intelligent classification and identification technology of space targets in the field of deep space exploration.

In this paper, a foreign object detection (FOD) system based on thermal infrared imaging is designed, taking advantage of its benefits, such as strong penetration in extreme weather conditions, high accuracy of non-contact detection, and no side effect on the wireless power transfer (WPT) system. Firstly, theoretical analysis and improvements are made on image processing techniques such as bilinear interpolation, Gaussian filtering, Operational Test Support Unit (Ostu) adaptive binarization, and morphological operations. Next, the processing advantages of the enhanced algorithms are confirmed using the openCV toolbox. Then, the aforementioned image processing techniques are implemented using embedded programming via the STM32 microprocessor, and the system’s foreign object detection capability is also verified by designing experiments for the location and type of foreign object under 5kW power level. Finally, the experimental results show that the WPT foreign object detection system is effective in the case of detecting and locating foreign objects that up to the size of a half yuan coin even an M4 nut, when the sensor is 15 cm away from the foreign object and the maximum viewing angle is within 110°. At this point, the minimum required temperature deviation between foreign object and background is 7°C, as well as the detection processing time is less than 0.5s, which fully satifies the demand for foreign obiect detection in the WPT system.

Image target recognition based on strapdown seeker is the key to extracting guidance information and realizing precise guidance. This paper proposes a new deep neural network target recognition framework. The framework extracts features from target template images and strapdown images at the front end and inputs the extracted feature maps into the deep neural network for target region recognition, thereby realizing the objects given by the template are accurately identified in the strapdown visible light seeker image. Aiming at the application scenario of strapdown image target recognition, this paper produced a target dataset, and tested typical algorithms such as SSD, YOLOv3, YOLOv4 and YOLOv5 on the proposed framework. The results show that the proposed framework can effectively recognize the objects given by the template in strapdown images.