Average Frame Research Articles

Development of an Optronic Aiming System for Target Tracking on the S60 57mm Cannon Weapon Control System Using a Camera

The development of technological science towards the defense of a country is growing rapidly. The country needs increasingly advanced defense technology but is constrained by an increasingly large budget due to dependence on producing countries. The state provides opportunities for technocrats to carry out research that can later create a defense technology that is inexpensive and does not burden the state budget. The artificial S60 57mm cannon is one of the cannons that functions as an air attack deterrent. The operating system of this weapon is still manual. This research will design and prototype an automatic Weapon Control System on the S60 57mm gun. In this development using the Atmega 8 microcontroller is the controller and interface for transferring data from the camera (Optronic) to the servo motor. The result of the camera position (optronics) will be followed in real-time by the gun barrel. In designing a target tracking control system automatically using the proportional control method. The results of testing the ability of the servo motor as a camera driver to follow targets or moving objects with a maximum angular speed of 15.5 degrees/second at a speed of 0.3 seconds, the average frame rate of the camera are 60 fps. c. Based on testing the servo motor using the Atmega8 microcontroller support, the reading of the angle direction of the servo motor is displayed in the Delphi software. Based on the test results, there is a difference between mathematical calculations and measurements using an arc ruler. With mathematical calculations, there is a difference with a value of 1.06 degrees per 1 degree.

Lightweight tomato real-time detection method based on improved YOLO and mobile deployment

The current deep-learning-based tomato target detection algorithm has many parameters; it has drawbacks of large computation, long time consumption, and reliance on high-computing-power devices such as graphics processing units (GPU). In this study, we propose a lightweight improved YOLOv5 (You Only Look Once) based algorithm to achieve real-time localization and ripeness detection of tomato fruits. Initially, this algorithm used a down-sampling convolutional layer instead of the original focus layer, reconstructing the backbone network of YOLOv5 using the bneck module of MobileNetV3. Then, it performs channel pruning for the neck layer to further reduce the model size and uses a genetic algorithm for hyperparameter optimization to improve detection accuracy. We evaluate the improved algorithm using a homemade tomato dataset. The experimental results demonstrated that the improved model number of parameters and floating point operations per second (FLOPs) were compressed by 78% and 84.15% compared to the original YOLOv5s, while the mAP reached 0.969. Meanwhile, the detection speed on the central processing unit (CPU) platform was 42.5 ms, which was 64.88% better. This study further utilized the Nihui convolutional neural network (NCNN) framework to quantize the improved model and developed an Android-based real-time tomato monitoring application (app). Experimental results demonstrated that the 16-bit quantized model achieved an average detection frame rate of 26.5 frames per second (fps) on the mobile side with lower arithmetic power, which was 268% better than the original YOLOv5s, and the model size was reduced by 51.1% while achieving a 93% true detection rate.

Cows' legs tracking and lameness detection in dairy cattle using video analysis and Siamese neural networks

The analysis of a cow’s motion behavior is based on the positioning and tracking of its legs. This paper proposes a Siamese attention model (Siam-AM) fusion attention mechanism to obtain the automatic tracking and monitoring of cow legs in large-scale farms. Firstly, the features of the cow’s leg are extracted using the first frame of motion image data. The search area in the subsequent frames is also input into the network to extract additional features. Then an attention module is used to assign weights to the features. The similarity between different regions and image features of the first frame was compared. The image with the most significant similarity was taken as the predicted position of the cow’s leg. Finally, the relative step size of the cow’s front and rear legs is calculated using the leg’s position coordinates. A Support Vector Machine (SVM) classifier is constructed to realize the lameness detection of the cow. The experimental results show that the proposed Siam-AM algorithm has an average tracking accuracy of 93.80 %, an average frame rate of 57 fps, and a lameness detection accuracy of 94.73 %. These results provide an effective method for accurate tracking and lameness detection of a cow’s leg.

Retracted] A Voice Recognition Sensor and Voice Control System in an Intelligent Toy Robot System

In order to make the toy robot more entertaining, interesting, and intelligent, a voice recognition sensor and voice control system in the intelligent toy robot system are proposed. The system builds an overall system architecture including a client and a server. Through the camera calibration and data transmission module of the client, it collects images and calculates the internal and external parameters of the camera and transmits the image and external parameter data to the server. With the images and external parameter data transmitted from the terminal, a background image is constructed and the camera position and angle are updated in real time to complete the fusion of virtual and real scenes. Through the motion control part of the user interaction module, hearing‐impaired children can control the movement and rotation of smart toys. The experimental results show that the system has high communication synchronization and stability and can realize high‐precision control of smart toys, and the average frame rate can reach 30.97 f/s. The beneficial effect of the system is that it has various functions, has the effect of speech recognition, and is highly interesting.

Visual Object Tracking Based on Modified LeNet-5 and RCCF

The field of object tracking has recently made significant progress. Particularly, the performance results in both deep learning and correlation filters, based trackers achieved effective tracking performance. Moreover, there are still some difficulties with object tracking for example illumination and deformation (DEF). The precision and accuracy of tracking algorithms suffer from the effects of such occurrences. For this situation, finding a solution is important. This research proposes a new tracking algorithm to handle this problem. The features are extracted by using Modified LeNet-5, and the precision and accuracy are improved by developing the Real-Time Cross-modality Correlation Filtering method (RCCF). In Modified LeNet-5, the visual tracking performance is improved by adjusting the number and size of the convolution kernels in the pooling and convolution layers. The high-level, middle-level, and handcraft features are extracted from the modified LeNet-5 network. The handcraft features are used to determine the specific location of the target because the handcraft features contain more spatial information regarding the visual object. The LeNet features are more suitable for a target appearance change in object tracking. Extensive experiments were conducted by the Object Tracking Benchmarking (OTB) databases like OTB50 and OTB100. The experimental results reveal that the proposed tracker outperforms other state-of-the-art trackers under different problems. The experimental simulation is carried out in python. The overall success rate and precision of the proposed algorithm are 93.8% and 92.5%. The average running frame rate reaches 42 frames per second, which can meet the real-time requirements.

Evaluation of WebRTC in the Cloud for Surgical Simulations: A case study on Virtual Rotator Cuff Arthroscopic Skill Trainer (ViRCAST).

Web Real-Time Communication (WebRTC) is an open-source technology which enables remote peer-to-peer video and audio connection. It has quickly become the new standard for real-time communications over the web and is commonly used as a video conferencing platform. In this study, we present a different application domain which may greatly benefit from WebRTC technology, that is virtual reality (VR) based surgical simulations. Virtual Rotator Cuff Arthroscopic Skill Trainer (ViRCAST) is our testing platform that we completed preliminary feasibility studies for WebRTC. Since the elasticity of cloud computing provides the ability to meet possible future hardware/software requirements and demand growth, ViRCAST is deployed in a cloud environment. Additionally, in order to have plausible simulations and interactions, any VR-based surgery simulator must have haptic feedback. Therefore, we implemented an interface to WebRTC for integrating haptic devices. We tested ViRCAST on Google cloud through haptic-integrated WebRTC at various client configurations. Our experiments showed that WebRTC with cloud and haptic integrations is a feasible solution for VR-based surgery simulators. From our experiments, the WebRTC integrated simulation produced an average frame rate of 33 fps, and the hardware integration produced an average lag of 0.7 milliseconds in real-time.

Real-Time Tracking Based on Improved YOLOv5 Detection in Orchard Environment for Dragon Fruit

Highlights This method has achieved faster detection speed while maintaining accuracy. It is a real-time tracking method that can track dragon fruits in orchard environments in real-time. The introduction of an attention mechanism in the network provides good robustness to changes in lighting and target scale. Abstract. This article addresses the issue of dragon fruit real-time detection in orchard environments and proposes a real-time detection and tracking model for dragon fruit using an improved YOLOv5 object detection algorithm and Deep-sort object tracking algorithm. By applying real-time tracking to dragon fruit harvesting, the tracking algorithm provides timely feedback on the fruit's location, allowing for prompt correction of environmental issues that may affect the accuracy of the harvesting process. This approach enhances the robustness of the target positioning algorithm. First,based on the YOLOv5 object detection algorithm, the Convolutional Block Attention Module and Transformer self-attention mechanism are introduced to construct a YOLOv5s-DFT object detection model that is more suitable for dragon fruit detection. Next, Combining the Deep-sort multi-object tracking algorithm, this article proposes a real-time detection and tracking method for dragon fruit in the orchard environment. The YOLOv5s-DFT model was trained and experimented with using a self-built dataset. The trained model weight is only 19.26% of YOLOv7. The experimental result shows that, while ensuring detection accuracy, YOLOv5s-DFT has a faster detection speed in dragon fruit detection, with an average frame time of 0.01673 s, which is 0.00422 s faster than the original YOLOv5s. When tracking dragon fruit using the Deep-sort tracking algorithm, it can track dragon fruit at a speed of 47.08 frames per second. When utilizing the Deep-sort tracking algorithm to track dragon fruit, it achieves a tracking speed of 47.08 frames per second, enabling real-time acquisition of the fruit's position information. This technology provides technical assistance for the intelligent harvesting of dragon fruit and the intelligent management of dragon fruit orchards. Keywords: Dragon fruit, Improved YOLOv5, Orchard environment, Real-time tracking.

Proposta de novo sistema de escore de estrutura corporal para bovinos Nelore

Abstract A new frame score system was developed for Nellore cattle, based upon ideal slaughter weights to achieve desired degrees of carcass fat cover. Data consisting of 688 complete records including sex, age, weight, hip height, and subcutaneous fat on growing Nellore cattle (190 males and 498 females) were obtained from the National Association of Breeders and Researchers (ANCP), Ribeirão Preto – SP, Brazil. Weight data were regressed on sex, age, hip height, and fat cover as covariates. Age was then fixed at 550 days and fat cover was set to 6 mm, and new equations for frame scores (1 to 11) in Nellore cattle were developed: FrameNellore (males) = -42.43 + 0.04919 × Age + 0.3368 × Height – 0.0003369 × Age × Height and FrameNellore (females) = -41.76 + 0.04919 × Age + 0.3368 × Height – 0.0003369 × Age × Height (Age and Height in days and cm, respectively). The Nellore equations resulted in average frame of 5.3, ranging from 2.3 to 7.7. Values were similar for males (5.4) and females (5.3). By contrast, previous equations gave higher values, that were very different between males and females. The new Nellore frame scores appear reasonable, particularly regarding body composition.

Design and Development of a Multi-Sided Tabletop Augmented Reality 3D Display Coupled with Remote 3D Imaging Module

This paper proposes a tabletop augmented reality (AR) 3D display paired with a remote 3D image capture setup that can provide three-dimensional AR visualization of remote objects or persons in real-time. The front-side view is presented in stereo-3D format, while the left-side and right-side views are visualized in 2D format. Transparent glass surfaces are used to demonstrate the volumetric 3D augmentation of the captured object. The developed AR display prototype mainly consists of four 40 × 30 cm2 LCD panels, 54% partially reflective glass, an in-house developed housing assembly, and a processing unit. The capture setup consists of four 720p cameras to capture the front-side stereo view and both the left- and right-side views. The real-time remote operation is demonstrated by connecting the display and imaging units through the Internet. Various system characteristics, such as range of viewing angle, stereo crosstalk, polarization perseverance, frame rate, and amount of reflected and transmitted light through partially reflective glass, were examined. The demonstrated system provided 35% optical transparency and less than 4% stereo crosstalk within a viewing angle of ±20 degrees. An average frame rate of 7.5 frames per second was achieved when the resolution per view was 240 × 240 pixels.

DSMFFNet: Depthwise separable multiscale feature fusion network for bridge detection in very high resolution satellite images

Bridge detection methods based on deep learning have many parameters, complex calculations, and serious errors and missed detections for multiscale bridges. To solve the above problems, a depth-wise separable multiscale feature fusion network (DSMFFNet) is proposed for efficient and accurate bridge detection in very high resolution satellite images (VHR). First, depth-wise separable convolution was used to build a backbone feature extraction network to extract the bridge features. Second, to better match bridges of different scales, multiscale receptive fields were obtained by multibranch parallel dilated convolution at the last layer of the feature map. Then, to make full use of the details and semantic information of the bridges at different depths, the three effective feature layers of the bridges at different levels are fused by a multiscale feature pyramid. The experimental results showed that the mean average precision (mAP) and frame per second (FPS) of the proposed method reach 94.26% and 60.04%, which can lead most of the mainstream object detection networks in accuracy and speed and can be integrated into the mobile end to complete the task of high-precision and fast bridge detection.

Social Distancing Monitoring System using Deep Learning

COVID-19 has been declared a pandemic in the world by 2020. One way to prevent COVID-19 disease, as the World Health Organization (WHO) suggests, is to keep a distance from other people. It is advised to stay at least 1 meter away from others, even if they do not appear to be sick. The reason is that people can also be the virus carrier without having any symptoms. Thus, many countries have enforced the rules of social distancing in their Standard Operating Procedure (SOP) to prevent the virus spread. Monitoring the social distance is challenging as this requires authorities to carefully observe the social distancing of every single person in a surrounding, especially in crowded places. Real-time object detection can be proposed to improve the efficiency in monitoring the social distance SOP inspection. Therefore, in this paper, object detection using a deep neural network is proposed to help the authorities monitor social distancing even in crowded places. The proposed system uses the You Only Look Once (YOLO) v4 object detection models for the detection. The proposed system is tested on the MS COCO image dataset with a total of 330,000 images. The performance of mean average precision (mAP) accuracy and frame per second (FPS) of the proposed object detection is compared with Faster Region-based Convolutional Neural Network (R-CNN) and Multibox Single Shot Detector (SSD) model. Finally, the result is analyzed among all the models.

Automatic tunnel lining crack detection via deep learning with generative adversarial network-based data augmentation

Aiming at solving the challenges of insufficient data samples and low detection efficiency in tunnel lining crack detection methods based on deep learning, a novel detection approach for tunnel lining crack was proposed, which is based on pruned You Look Only Once v4 (YOLOv4) and Wasserstein Generative Adversarial Network enhanced by Residual Block and Efficient Channel Attention Module (WGAN-RE). In this study, a data augmentation method named WGAN-RE was proposed, which can achieve the automatic generation of crack images to enrich data set. Furthermore, YOLOv4 was selected as the basic model for training, and a pruning algorithm was introduced to lighten the model size, thereby effectively improving the detection speed. Average Precision (AP), F1 Score (F1), model size, and Frames Per Second (FPS) were selected as evaluation indexes of the model performance. Results indicate that the storage space of the pruned YOLOv4 model is only 49.16 MB, which is 80% compressed compared with the model before pruning. In addition, the FPS of the model reaches 40.58f/s, which provides a basis for the real-time detection of tunnel lining cracks. Findings also demonstrate that the F1 score and AP of the pruned YOLOv4 are only 0.77% and 0.50% lower than that before pruning, respectively. Besides, the pruned YOLOv4 is superior in both model accuracy and efficiency compared with YOLOv3, SSD, and Faster RCNN, which indicated that the pruned YOLOv4 model can realize the accurate, fast and intelligent detection of tunnel lining cracks in practical tunnel engineering.

Long video-based action segmentation for earthmoving excavators using improved Temporal Convolutional Network models

Earthwork excavator, as an all-terrain and high-efficiency earthwork excavation equipment, has been widely used in earthwork sites. It is very necessary to analyze the work of earthmoving excavator by means of machine vision. In this paper, the action segmentation method based on long video was applied to the analysis and recogniton of the excavator’s action, and compared with other two current best action segmentation models using the real construction site video. Firstly, the sequence features of the excavator’s work video obtained at the cnstruction site was extracted through 3D convolution method, and then two differernt networks with the extracted sequence features were trained and tested. The experimental results showed that the average frame accuracy of MS-TCN model and ASRF model in excavator action segmentation were 82.6490% and 86.1042% respectively. However, for the recognition task under different working environment, the performance of the two models is quite different. The experimental results manifest that the motion segmentation model based on long video reached good results in excavator motion recognition in earthmoving operation. And it is helpful to analyze the long video working behavior sequence of excavator. This research contributes to the identification of critical elements that explains serial action and to the development of a new application scenario for vision-based behavior segmentation network. Additionally, the results of this study were helpful to automatically analyze the working efficiency and monitor the productivity of earthmoving excavators. Using this kind of data-driven decision can improve the work efficiency of earthmoving excavator and promote the project progress.

Goat-Face Recognition in Natural Environments Using the Improved YOLOv4 Algorithm

In view of the low accuracy and slow speed of goat-face recognition in real breeding environments, dairy goats were taken as the research objects, and video frames were used as the data sources. An improved YOLOv4 goat-face-recognition model was proposed to improve the detection accuracy; the original backbone network was replaced by a lightweight GhostNet feature extraction network. The pyramid network of the model was improved to a channel management mechanism with a spatial pyramid structure. The path aggregation network of the model was improved into a fusion network with residual structure in the form of double parameters, in order to improve the model’s ability to detect fine-grained features and distinguish differences between similar faces. The transfer learning pre-training weight loading method was adopted, and the detection speed, the model weight, and the mean average precision (mAP) were used as the main evaluation indicators of the network model. A total of 2522 images from 30 dairy goats were augmented, and the training set, validation set, and test set were divided according to 7:1:2. The test results of the improved YOLOv4 model showed that the mAP reached 96.7%, and the average frame rate reached 28 frames/s in the frontal face detection. Compared with the traditional YOLOv4, the mAP improved by 2.1%, and the average frame rate improved by 2 frames/s. The new model can effectively extract the facial features of dairy goats, which improves the detection accuracy and speed. In terms of profile face detection, the average detection accuracy of the improved YOLOv4 goat-face-recognition network can reach 78%. Compared with the traditional YOLOv4 model, the mAP increased by 7%, which effectively demonstrated the improved profile recognition accuracy of the model. In addition, the improved model is conducive to improving the recognition accuracy of the facial poses of goats from different angles, and provides a technical basis and reference for establishing a goat-face-recognition model in complex situations.

Deep Learning for Clothing Style Recognition Using YOLOv5.

With the rapid development of artificial intelligence, much more attention has been paid to deep learning. However, as the complexity of learning algorithms increases, the needs of computation power of hardware facilities become more crucial. Instead of the focus being on computing devices like GPU computers, a lightweight learning algorithm could be the answer for this problem. Cross-domain applications of deep learning have attracted great interest amongst researchers in academia and industries. For beginners who do not have enough support with software and hardware, an open-source development environment is very helpful. In this paper, a relatively lightweight algorithm YOLOv5s is addressed, and the Google Colab is used for model training and testing. Based on the developed environment, many state-of-art learning algorithms can be studied for performance comparisons. To highlight the benefits of one-stage object detection algorithms, the recognition of clothing styles is investigated. The image samples are selected from datasets of fashion clothes and the web crawling of online stores. The image data are categorized into five groups: plaid; plain; block; horizontal; and vertical. Average precison, mean average precison, recall, F1-score, model size, and frame per second are the metrics used for performance validations. From the experimental outcomes, it shows that YOLOv5s is better than other learning algorithms in the recognition accuracy and detection speed.

Predictors of shear wave propagation speed assessed by shear wave elastography

Abstract Background Cardiac shear wave elastography (SWE) is a novel ultrasound-based method that assesses the propagation speed of shear waves travelling in the myocardium induced by e.g. mitral valve closure (MVC). The propagation speed of these waves is related to the stiffness of the myocardium. The capability of SWE to evaluate myocardial stiffness has already been demonstrated in several publications. However, which factors beside the intrinsic mechanical properties of the myocardium influence shear wave speed in vivo has not been extensively investigated. Purpose The aim of this study was to investigate the influence of clinical parameters as well as echocardiographic indicators of myocardial tissue properties and left ventricular (LV) filling pressures on shear wave propagation speed after MVC. Methods 219 subjects (59±17 years; 140 males) were included in the study. Of those, 49 were healthy volunteers, 25 were patients with cardiac amyloidosis, 42 with hypertrophic cardiomyopathy, 35 with hypertensive heart disease and 68 with coronary artery disease. Subjects were scanned with an experimental ultrasound scanner using diverging wave imaging to acquire images at a high temporal resolution (average frame rate: 1167±414 Hz). Shear waves after MVC were visualized on M-mode maps along the interventricular septum which were colour coded for tissue acceleration (Figure 1). The propagation speed was calculated by semi-automatically measuring the spatiotemporal slope of the shear wave. Univariate and multivariate linear regression analysis was performed to identify variables associated with shear wave propagation speed. Results Univariate regression analysis revealed an association of the following variables with shear wave speed: age, BMI, systolic blood pressure, E wave, average e', E/e', left atrial volume index (LAVI), grade of diastolic dysfunction, isovolumic relaxation time, septal wall thickness, LV wall mass and presence of cardiomyopathy. These variables were added to a multivariate model. Predictors of shear wave speed after MVC were BMI, E wave, average e', LAVI, LV wall mass and presence of cardiomyopathy (R2=0.53) (Table 1). Conclusions The results of this study indicate that both myocardial structural properties (reflected by predictors LV wall mass and presence of cardiomyopathy) and LV filling pressures (reflected by predictors E wave, average e' and LAVI) affect shear wave speed. These findings suggest that SWE has the potential to assess structural as well as functional changes to the LV, which should be further explored. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Research foundation Flanders - FWO

Accurate detection of dairy cow mastitis with deep learning technology: a new and comprehensive detection method based on infrared thermal images

Mastitis is one of the most common diseases in dairy cows and has a negative impact on their welfare and life, causing significant economic losses to the dairy industry. Many attempts have been made to develop a detection method for mastitis using thermal infrared thermography. However, the use of this detection technique to determine the health of the cow's udder is susceptible to external factors, resulting in inaccurate detection of dairy cow mastitis. Therefore, this study explored a new and comprehensive detection method of dairy cow mastitis based on infrared thermal images. This method combined the left and right udder skin surface temperature (USST) difference detection method with the ocular surface temperature and USST difference detection method with improvements. The effect of external factors on dairy cow USST was effectively reduced. In addition, after comparing different target localisation algorithms, this paper used the You Only Look Once v5 (YOLOv5) deep learning network model to obtain the temperature information of eyes and udders, and mastitis detection of dairy cows was performed. A total of 105 dairy cows passing through a passage were randomly selected from the thermal infrared video and detected by the new and comprehensive detection method, and the results of cow mastitis detection were compared with somatic cell count. The results showed that the accuracy, specificity, and sensitivity of mastitis detection were 87.62, 84.62, and 96.30%, respectively. Using the YOLOv5 deep learning network model to locate the key parts of the cow had a good effect, with an average accuracy of 96.1%, and an average frame rate of 116.3f/s. The detection accuracy of dairy cow mastitis by deep learning technology combined with the detection method in this paper reached 85.71%. The results showed that the new and comprehensive detection method based on infrared thermal images can be used for the detection of dairy cow mastitis with high detection accuracy. This method can reduce the influence of external factors and can be integrated into the automatic identification system of dairy mastitis based on YOLOv5 to realise on-site monitoring of dairy mastitis.

Definition of electromagnetic local spatial densities for composite spin- 1/2 systems

An unambiguous definition of the electromagnetic spatial densities for a spin-1/2 system is proposed and worked out in the zero average momentum frame and in moving frames. The obtained results are compared with the traditional definition of the densities in terms of the three-dimensional Fourier transforms of the electromagnetic form factors in the Breit frame.

Lite-3DCNN Combined with Attention Mechanism for Complex Human Movement Recognition.

Three-dimensional convolutional network (3DCNN) is an essential field of motion recognition research. The research work of this paper optimizes the traditional three-dimensional convolution network, introduces the self-attention mechanism, and proposes a new network model to analyze and process complex human motion videos. In this study, the average frame skipping sampling and scaling and the one-hot encoding are used for data pre-processing to retain more features in the limited data. The experimental results show that this paper innovatively designs a lightweight three-dimensional convolutional network combined with an attention mechanism framework, and the number of parameters of the model is reduced by more than 90% to only about 1.7 million. This study compared the performance of different models in different classifications and found that the model proposed in this study performed well in complex human motion video classification. Its recognition rate increased by 1%–8% compared with the C3D model.

Real‐time recognition and warning of mask wearing based on improved YOLOv5 R6.1

Since the new crown epidemic, mask-wearing has become a new normal in people's work and life. The inspection mechanism for mask-wearing at the entrance and exit of public places is seriously insufficient. The phenomenon of “pick-up on entry” has led to the severe formalization of mask-wearing inspection. Manual detection of mask-wearing in an open and dynamic crowded environment is unrealistic, which is not only time-consuming and labor-intensive but also cannot achieve early warning throughout the entire process. In response to this problem, this paper proposes a real-time recognition and early warning method for mask-wearing in an open, dynamic, complex environment based on improved YOLOv5 R6.1. First, replacing the first Conv structure of the backbone network in the YOLOv5 R6.1 model with an improved Stem structure to minimize the computational overhead while improving the performance. Then by normalizing the data, the random erasure data expansion technique is used to enhance the antiocclusion robustness of the algorithm. Finally, according to the mask-wearing specification in the training data set, optimizing and adjusting the anchor box parameters of the YOLOv5 R6.1 model to improve the model's ability to recognize small targets. The experiments are based on open data sets, and the results show that the mean precision (mAP), precision, and recall of this method reach 92.9%, 94.1%, and 88.5% on average, and the average frames per second (FPS) reaches 117. Moreover, the mAP and FPS are improved by an average of 6.5% and 474% compared with algorithms based on RetinaNet, Attention-Retina, Single Shot multibox Detector, Fast-RCNN, YOLOv4, and YOLOv5.