Requirements Of Real-time Detection Research Articles

Insulator detection and damage identification based on improved lightweight YOLOv4 network

Aiming at the problem that the detection and damage recognition algorithm of insulator aerial images is difficult to be applied to the embedded platform, this paper proposed an improved Tiny-YOLOv4 lightweight target detection network algorithm. This algorithm combined the Self-Attention mechanism and ECA-Net (Efficient Channel Attention Neural Networks), which could greatly reduce the complexity of the original YOLOv4 algorithm, and the model size is 24.9 MB. Under the condition of ensuring the detection accuracy (>91%), the detection speed is as high as 94FPS. It is transplanted to the Jetson Xavier NX embedded platform, and the average detection speed reaches 22FPS, which effectively meets the real-time detection requirements in the power inspection.

Motion Target Detection and Recognition Based on YOLOv4 Algorithm

With the application of artificial intelligence more and more widely, the target detection of artificial intelligence “eyes” is becoming more and more important, which can give the machine the ability to detect and recognize image content in the environment. In this paper, YOLOv4 deep learning motion target detection algorithm is used to achieve the positioning and recognition of motion targets, to identify and mark the location and type of objects contained in the image, to achieve the detection of pictures, videos and camera realtime image detection, so that the machine has the most basic visual ability. Using YOLOv4 algorithm, the speed requirement of real-time detection can be realized with certain precision, and the model mAP value reached 69.48%.

A Real-Time Detection Algorithm for Kiwifruit Defects Based on YOLOv5

Defect detection is the most important step in the postpartum reprocessing of kiwifruit. However, there are some small defects difficult to detect. The accuracy and speed of existing detection algorithms are difficult to meet the requirements of real-time detection. For solving these problems, we developed a defect detection model based on YOLOv5, which is able to detect defects accurately and at a fast speed. The main contributions of this research are as follows: (1) a small object detection layer is added to improve the model’s ability to detect small defects; (2) we pay attention to the importance of different channels by embedding SELayer; (3) the loss function CIoU is introduced to make the regression more accurate; (4) under the prerequisite of no increase in training cost, we train our model based on transfer learning and use the CosineAnnealing algorithm to improve the effect. The results of the experiment show that the overall performance of the improved network YOLOv5-Ours is better than the original and mainstream detection algorithms. The mAP@0.5 of YOLOv5-Ours has reached 94.7%, which was an improvement of nearly 9%, compared to the original algorithm. Our model only takes 0.1 s to detect a single image, which proves the effectiveness of the model. Therefore, YOLOv5-Ours can well meet the requirements of real-time detection and provides a robust strategy for the kiwi flaw detection system.

Object detection algorithm based on feature enhancement

Recently, many excellent algorithms have made great progress in object detection, but there are also problems in these algorithms’ performance on targets of different sizes, and in particular in small object detection. Aiming at the problem of insufficient feature representation by the feature extractor, in this paper we propose a lightweight algorithm to improve feature extraction. The algorithm includes three modules. First, considering that the shallow features in feature extraction contain much background noise, in this paper we design a multi-level feedback propagation model based on a Gaussian high-pass filter. The shallow layers are enhanced using the filter and then back-propagated to add the upper shallow layer features and obtain new shallow layer features. This process is performed on the newly generated shallow layer for n iterations, which is beneficial for enhancing targets in the foreground area and suppressing background noise. Second, we form a stacked dilated convolution module with different dilation rates to cover the entire deep feature layer densely, which enlarges the receptive field and enriches the contextual information. Finally, we build a multi-scale fusion module to fuse the above-mentioned enhanced shallow and deep features to obtain output features with powerful representational ability for detection tasks. In addition, the model is easily embedded into existing approaches to enhance their performance. We build the model on the VGG-16 and ResNet-50 backbones and successfully applied it on Darknet-19 and Darknet-53 to verify its effectiveness and stability. The experiments on the COCO dataset prove that the proposed algorithm outperforms the state-of-art methods, with a mean average precision improvement reaching 2% on average. The effect is remarkable on small targets and complex backgrounds. Furthermore, it does not affect the detection speed significantly, so real time detection requirements can still be met.

Vehicle and Pedestrian Detection Based on Improved YOLOv4-tiny Model

Target detection is the basic technology of automatic driving system. Deep learning has gradually become the mainstream target detection algorithm because of its powerful feature extraction ability and adaptive ability. How to ensure accuracy and speed is a great challenge in the field of target detection. In order to solve the problems of high miss detection rate of small target and difficult to realize embedded real-time detection in the process of complex environment detection by deep learning method, this paper adds two auxiliary remaining network blocks in the backbone network. So that the backbone network can extract the global and local features of the detected object, and carry out feature extraction based on the feature pyramid network Fusion, adding a scale to form a three scale prediction, to improve the problem of poor detection accuracy of yolov4-tiny network. The simulation results show that: Compared with yolov4-tiny, the accuracy of the improved network structure is improved by 3.3%, and the detection speed is 251 fps, which ensures the requirements of real-time detection. This algorithm has good detection effect in the case of lack of illumination and target occlusion, and its detection accuracy on the mixed data set is better than that of the contrast algorithm, which meets the real-time detection conditions and is suitable for deployment on the embedded system carried by the car.

Transcription-Factor-based Biosensor Engineering for Applications in Synthetic Biology.

Transcription-factor-based biosensors (TFBs) are often used for metabolite detection, adaptive evolution, and metabolic flux control. However, designing TFBs with superior performance for applications in synthetic biology remains challenging. Specifically, natural TFBs often do not meet real-time detection requirements owing to their slow response times and inappropriate dynamic ranges, detection ranges, sensitivity, and selectivity. Furthermore, designing and optimizing complex dynamic regulation networks is time-consuming and labor-intensive. This Review highlights TFB-based applications and recent engineering strategies ranging from traditional trial-and-error approaches to novel computer-model-based rational design approaches. The limitations of the applications and these engineering strategies are additionally reviewed.

Infrared target detection based on Gaussian model and Hungarian algorithm

ABSTRACT Based on Gaussian distribution, a method for modelling targets is proposed. The trajectories of infrared targets are predicted by Kalman filtering. The Munkres version of the Hungarian algorithm is utilized to find the optimum assignment between the trajectories and the positions of real targets. Finally, based on the open dataset, we prove that the proposed algorithm is able to find and track infrared dim targets more effectively, while satisfying the requirements of real-time detection.

Real-time detection method of driver fatigue state based on deep learning of face video

The use of face video information for driver fatigue detection has received extensive attention because of its low cost and non-invasiveness. However, the current vehicle-mounted embedded device has insufficient memory and limited computing power, which cannot complete the real-time detection of driver fatigue based on deep learning. Therefore, this paper designs a lightweight neural network model to solve this problem. The model includes object detection and fatigue detection. First, a lightweight object detection network is designed, which can quickly identify the opening and closing states of the driver’s eyes and mouth in the time series video. Secondly, the EYE-MOUTH (EM) driver fatigue detection model is designed, which encodes the driver’s eye and mouth opening and closing state, and calculates the driver’s PERCLOS (Percentage of Eyelid Closure over the Pupil) and FOM (Frequency of Open Mouth) according to the coding sequence. Finally, the multi-feature fusion judgment algorithm is used to realize the judgment of the driver’s fatigue state. The experimental results show that our method has an accuracy rate of 98.30% for drowsiness and yawning behaviors in a real vehicle environment, and a detection speed of 27FPS, which is better than other advanced methods and meets the requirements of real-time detection.

Event-Based Pedestrian Detection Using Dynamic Vision Sensors

Pedestrian detection has attracted great research attention in video surveillance, traffic statistics, and especially in autonomous driving. To date, almost all pedestrian detection solutions are derived from conventional framed-based image sensors with limited reaction speed and high data redundancy. Dynamic vision sensor (DVS), which is inspired by biological retinas, efficiently captures the visual information with sparse, asynchronous events rather than dense, synchronous frames. It can eliminate redundant data transmission and avoid motion blur or data leakage in high-speed imaging applications. However, it is usually impractical to directly apply the event streams to conventional object detection algorithms. For this issue, we first propose a novel event-to-frame conversion method by integrating the inherent characteristics of events more efficiently. Moreover, we design an improved feature extraction network that can reuse intermediate features to further reduce the computational effort. We evaluate the performance of our proposed method on a custom dataset containing multiple real-world pedestrian scenes. The results indicate that our proposed method raised its pedestrian detection accuracy by about 5.6–10.8%, and its detection speed is nearly 20% faster than previously reported methods. Furthermore, it can achieve a processing speed of about 26 FPS and an AP of 87.43% when implanted on a single CPU so that it fully meets the requirement of real-time detection.

An Algorithm for Detecting the Integrity of Outer Frame Protection Net on Construction Site Based on Improved SSD

In recent years, the continuous development of the basic construction industry has promoted the progress of urban social economy, and the safety issues in construction have attracted more and more attention. The integrity of the external protection net is of great significance to the safety of the construction personnel. Most of the traditional methods of detecting damage to the protection net have problems such as poor real-time performance and low detection accuracy. In this paper, we present a novel approach based on SSD for solving the problem of the integrity detection of the external frame protection net in the construction site. The improved network merges adjacent feature layers to enrich the semantics of shallow feature maps, thereby improving the detection accuracy of the network, and simultaneously introduces a shallower feature layer and a more advanced receptive field module to improve the detection ability of small targets. The experimental results show that compared with the SSD algorithm, the improved algorithm has higher detection accuracy in the detection task of the external protective net on the construction site, and can meet the real-time detection requirements.

Target motion detection algorithm based on dynamic threshold

Realizing moving target detection through visual algorithms is a major branch of computer technology. Moving target detection has a wide range of applications in many fields such as video surveillance, unmanned driving, and behavior research. The current algorithm has been able to achieve moving target detection well in a static background, but the detection of multiple moving targets in a dynamic background is still a challenging research area. The main pain points are complex background, light changes, noise interference and Occlusion and real-time detection requirements, etc. Aiming at the detection of moving targets in complex scenes, a moving target detection algorithm based on dynamic threshold is proposed. The algorithm first uses the gray difference gradient of three adjacent frames of the video sequence to calculate the segmentation threshold of foreground moving targets to realize the dynamic threshold. Then use dynamic thresholds to achieve multi-target motion detection in complex backgrounds or scenes with changing lighting. Because the gradient change of adjacent frames of the scene is considered in the process of moving target detection and segmentation, and the gradient change is converted into a dynamic threshold to participate in moving target detection and segmentation, the algorithm can adapt to scenes with changing lighting. Experiments show that this algorithm is feasible for moving target detection in complex scenes, and has a lower error rate than that of traditional algorithms, and is more suitable for to complex or changing lighting scenes.

RTED-SD: A Real-Time Edge Detection Scheme for Sybil DDoS in the Internet of Vehicles

As the increasing popularity and development of the Internet of Vehicles (IoV), it is important to ensure the reliability and safety for IoV. However, communicating in an open-access environment makes road safety, communication security and privacy issues facing great challenges. Also, the high requirement of real-time detection and reponse to the security issues makes edge detection a more and more important research subject. Among all the challenges, Sybil Denial of Service (DoS) Attack is one of the most severe threats to the IoV safety, which can lead to traffic jams and other safety issues. In this paper, we introduce a Real-Time Edge Detection Scheme for Sybil DDoS in IoV. We use the entropy theory to quantify the traffic distribution, and further design an algorithm named Fast Quartile Deviation Check (FQDC) to recognize and locate the attack. Furthermore, because of the calculation limit in IoV scenarios, we also optimize the calculation with some useful techniques, such as the optimized sliding window, the incremental calculation of entropy values and make it suitable for the IoV environment. Finally, we proposed a temporal index. Temporal False Omission Rate (TFOR), to measure the performance of response speed and omission rate. In our evaluation, we successfully detect all the Sybil DDoS attacks provided in the F2MD datasets, and have the average alarm delay of 4.9193 seconds, and the average TFOR of 1.6024%.

Fabric Defect Segmentation Method Based on Deep Learning

Fabric defect detection plays an essential role in the textile production process, which was widely applied in the textile industry. For fabric defect detection, many algorithms have been proposed. However, lots of important problems, such as the accuracy of detection, the computational complexity of the algorithm, and data imbalance, still needed to be addressed for application in industrial production. In this article, we propose an efficient convolutional neural network for defect segmentation and detection. The design of this framework significantly alleviates the manual annotation cost of the data set; it only needs few defect samples combined with standard samples to learn the potential feature of defects and obtain the location of defects with high accuracy. The network is divided into two parts: segmentation and decision. First, the fabric data set without training is utilized as the input of the segmentation network. Then, the output of the segmentation network is applied as the raw materials for training the decision network. Finally, a well-trained network is used to obtain the location of defects with high accuracy. The proposed method only demands almost 50 defect samples to get accurate segmentation results and can achieve the requirement of real-time detection with a speed of 25 frames per second (FPS). The experimental results based on a public data set and three self-made fabric data sets show that the proposed method significantly outperforms eight state-of-the-art methods in terms of accuracy and robustness.

3D M-Net: Object-Specific 3D Segmentation Network Based on a Single Projection.

The internal assembly correctness of industrial products directly affects their performance and service life. Industrial products are usually protected by opaque housing, so most internal detection methods are based on X-rays. Since the dense structural features of industrial products, it is challenging to detect the occluded parts only from projections. Limited by the data acquisition and reconstruction speeds, CT-based detection methods do not achieve real-time detection. To solve the above problems, we design an end-to-end single-projection 3D segmentation network. For a specific product, the network adopts a single projection as input to segment product components and output 3D segmentation results. In this study, the feasibility of the network was verified against data containing several typical assembly errors. The qualitative and quantitative results reveal that the segmentation results can meet industrial assembly real-time detection requirements and exhibit high robustness to noise and component occlusion.

Insulator Surface Breakage Recognition Based on Multiscale Residual Neural Network

Insulators are one of the widely used important devices in power systems, and proposing an accurate and effective online monitoring method for insulator defects is an urgent need for intelligent and informative power systems. We propose a multiscale residual neural network (MSResNN) for insulator surface breakage recognition. The proposed multiscale convolution filter uses three different size convolution kernels for convolution filtering and performs feature maps fusion, which can enrich the spatial correlation and channel correlation of feature maps. We investigated the effects of different learning rates on the model performance and confirmed that the model performs best with learning rate of $10^{-6}$ . Compared with VGG16, Inception_v3, and Resnet50, the MSResNN model has significant advantages in detection capability and occupies less computational resources, and the detection speed can meet the industrial real-time detection requirements. Through dark–light and noise experiments, it is demonstrated that the proposed method is robust and can effectively deal with low-quality insulator images.

Research on Intelligent Target Recognition Integrated With Knowledge

With the development of artificial intelligence technology, intelligent weapon systems that can automatically identify, lock on and strike targets have gradually appeared and can replace humans in executing simple decision-making commands. Target detection is a key part of intelligent weapons. At present, large-scale target detection has serious challenges such as long-tail data distributions, severe occlusion, and category ambiguity. The main detection algorithms only detect each independent area without considering the key semantic dependencies between objects. It has become a hot trend to apply deep learning to prior knowledge to form a model. This article uses both internal and external knowledge to instill a target detection system with human reasoning capabilities. Commonly used external embedded knowledge includes geometric relations, attributes, locations, etc. They have a common shortcoming in that they require large amounts of labeled data, and the integration costs are huge. The purpose of this article is to construct a general external prior knowledge module to guide network learning. By paying attention to the characteristics of each object in different semantic contexts, the characteristics of each object are adaptively enhanced, and the high-level semantics of all categories evolve on a global scale. Internal knowledge uses a convolutional attention module that can learn spatial and channel information at multiple scales. The experimental results show the superiority of our knowledge-YOLOv5. The proposed method achieved 1.7%, 2.2%, 1.1%, and 0.7% improvements over YOLOv5s, YOLOv5m, VOLOv51, and YOLOv5x, respectively, on the COCO data sets; and the proposed method also achieves a 0.9% improvement on the self-built data set. The trained lightweight model Knowledge-YOLOv5s is deployed on an NVIDIA Jetson TX2 through TensorRT acceleration, and the real-time detection frame is 20 ms, which meets the real-time detection requirements. This system can also be used as a module of an intelligent weapon system, which has certain referential significance for autonomous weapons and unmanned combat systems.

Aerial Dataset Integration For Vehicle Detection Based on YOLOv4

With the increasing application of UAVs in intelligent transportation systems, vehicle detection for aerial images has become an essential engineering technology and has academic research significance. In this paper, a vehicle detection method for aerial images based on the YOLOv4 deep learning algorithm is presented. At present, the most known datasets are VOC (The PASCAL Visual Object Classes Challenge), ImageNet, and COCO (Microsoft Common Objects in Context), which comply with the vehicle detection from UAV. An integrated dataset not only reflects its quantity and photo quality but also its diversity which affects the detection accuracy. The method integrates three public aerial image datasets VAID, UAVD, DOTA suitable for YOLOv4. The training model presents good test results especially for small objects, rotating objects, as well as compact and dense objects, and meets the real-time detection requirements. For future work, we will integrate one more aerial image dataset acquired by our lab to increase the number and diversity of training samples, at the same time, while meeting the real-time requirements.

Algorithm of Helmet Wearing Detection Based on AT-YOLO Deep Mode

The existing safety helmet detection methods are mainly based on one-stage object detection algorithms with high detection speed to reach the real-time detection requirements, but they can’t accurately detect small objects and objects with obstructions. Therefore, we propose a helmet detection algorithm based on the attention mechanism (AT-YOLO). First of all, a channel attention module is added to the YOLOv3 backbone network, which can adaptively calibrate the channel features of the direction to improve the feature utilization, and a spatial attention module is added to the neck of the YOLOv3 network to capture the correlation between any positions in the feature map so that to increase the receptive field of the network. Secondly, we use DIoU (Distance Intersection over Union) bounding box regression loss function, it not only improving the measurement of bounding box regression loss but also increases the normalized distance loss between the prediction boxes and the target boxes, which makes the network more accurate in detecting small objects and faster in convergence. Finally, we explore the training strategy of the network model, which improves network performance without increasing the inference cost. Experiments show that the mAP of the proposed method reaches 96.5%, and the detection speed can reach 27 fps. Compared with other existing methods, it has better performance in detection accuracy and speed.

Detection of heterogeneous parallel steganography for low bit-rate VoIP speech streams

This paper considers a new task of detecting heterogeneous parallel steganography (HPS) on streaming media. This task is to detect the existence of the confidential messages hidden in the frames of streaming media with multiple kinds of orthogonal steganographic methods. We target on detecting HPS in this work for low bit-rate Voice over Internet Protocol (VoIP) speech streams, which is a widely-used streaming medium. Specifically, two steganographic methods, i.e., Quantization Index Modulation and Pitch Modulation Steganography, are utilized to form the HPS. Detecting HPS on low bit-rate VoIP speech streams is challenging for existing steganalysis methods. To accomplish the target, we propose a novel deep model named as Steganalysis Feature Fusion Network (SFFN). SFFN consists of three sub-networks, i.e., a feature learning network, a feature fusion network and a classification network. With the three sub-networks, SFFN can effectively extract steganalysis features for the steganographic methods used in HPS and can fuse the features to make credible prediction. The experimental results demonstrate that our method is superior to the state-of-the-art steganalysis methods when detecting HPS. Besides, our method meets the requirement of real-time detection.

An intelligent target detection method of UAV swarms based on improved KM algorithm

Complete and efficient detection of unknown targets is the most popular application of UAV swarms. Under most situations, targets have directional characteristics so that they can only be successfully detected within specific angles. In such cases, how to coordinate UAVs and allocate optimal paths for them to efficiently detect all the targets is the primary issue to be solved. In this paper, an intelligent target detection method is proposed for UAV swarms to achieve real-time detection requirements. First, a target-feature-information-based disintegration method is built up to divide the search space into a set of cubes. Theoretically, when the cubes are traversed, all the targets can be detected. Then, a Kuhn-Munkres (KM)-algorithm-based path planning method is proposed for UAVs to traverse the cubes. Finally, to further improve search efficiency, a 3D real-time probability map is established over the search space which estimates the possibility of detecting new targets at each point. This map is adopted to modify the weights in KM algorithm, thereby optimizing the UAVs’ paths during the search process. Simulation results show that with the proposed method, all targets, with detection angle limitations, can be found by UAVs. Moreover, by implementing the 3D probability map, the search efficiency is improved by 23.4%–78.1%.