Wear Helmets Research Articles

Helmet and License Plate Detection

Abstract: Traffic accidents are one of the leading causes of death today. Motorcycle accidents lead to serious injuries. A helmet is important for every motorcyclist. However, many do not adhere to helmet laws. This is the software CNN uses to find motorcyclists without helmets. The structure consists of motorcycle detection, helmetless classification, and motorcycle licence plate recognition. The motorcycle is scanned with the feature vector HOG. Once the motorcycle is recognised by CNN, do motorcyclists wear helmets? Motorcycle licence plate if it is determined that the motorcyclist is not wearing a helmet. Motorcycles are recognised by Tesseract OCR.

Injury Pattern and Outcomes Following All-Terrain Vehicle Accidents in Kentucky Children: A Retrospective Study.

All-terrain vehicles (ATVs) pose a significant risk for morbidity and mortality amongst children. We hypothesize that current vague legislation regarding helmet use impacts injury patterns and outcomes in pediatric ATV accidents. The institutional trauma registry was queried for pediatric patients involved in ATV accidents from 2006 to 2019. Patient demographics and helmet wearing status were identified in addition to patient outcomes, such as injury pattern, injury severity score, mortality, length of stay, and discharge disposition. These elements were analyzed for statistical significance. 720 patients presented during the study period, which were predominantly male (71%, n = 511) and less than 16years old (76%, n = 543). Most patients were not wearing a helmet (82%, n = 589) at time of injury. Notably, there were 7 fatalities. A lack of helmet use is positively associated with head injury (42% vs 23%, P < .01), intracranial hemorrhage (15% vs 7%, P = .03), and associated with lower Glasgow Coma Scale (13.9 vs 14.4, P < .01). Children 16years and older were least likely to wear a helmet and most likely to incur injuries. Patients over 16years had longer lengths of stay, higher mortality, and higher need for rehabilitation. Not wearing a helmet is directly correlated with injury severity and concerning rates of head injury. Children 16years and older are at greatest risk for injury, but younger children are still at risk. Stricter state laws regarding helmet use are necessary to reduce pediatric ATV-related injury burden. level III retrospective comparative study.

A Survey of Electric-Scooter Riders’ Route Choice, Safety Perception, and Helmet Use

This study investigated electric-scooter (e-scooter) rider behaviors and preferences to inform ways to increase safety for e-scooter riders. Data was collected from 329 e-scooter riders via two online and one in-person survey. Survey questions considered rider roadway infrastructure preferences, safety perceptions, and helmet-wearing behavior. Protected bike lanes were more commonly indicated as the safest infrastructure (62.4%) but were less likely to be the most preferred infrastructure (49.7%). Sidewalks were better matched between riders, indicating them as their preferred riding infrastructure (22.7%) and the perceived safest riding infrastructure (24.5%). Riders had low feelings of safety and preference for riding on major/neighborhood streets or on unprotected bike lanes. Riders reported significant concern about being hit by a moving vehicle, running into a pothole/rough roadway, and running into a moving vehicle. In line with the Theory of Planned Behavior, a significant relationship was found between the frequency of riding and helmet-wearing behavior, with more frequent riders being more likely to wear helmets. Findings suggest that existing roadway infrastructure may pose safety challenges and encourage rider-selected workarounds. Public policy may consider emphasizing protected bicycle lane development, rather than helmet mandates, to support e-scooter riding safety for all vulnerable road users.

Design and Development of Automatic Detection System for Motercyclists without Helmet using Machine Learning

The use of two-wheelers as a mode of transport is increasing rapidly, but unfortunately, many riders neglect to wear helmets, which can lead to accidents and fatalities. To address this issue, many countries have implemented laws mandating the use of helmets for two-wheeler riders, and the police force often discourages this behavior by issuing traffic violation tickets. However, the process of issuing these tickets is often manual and tedious, which can lead to delays and errors. To solve this problem, a proposed system is to automate the process of detecting riders who are not wearing helmets. The system would use image processing algorithms to extract the license plate number of the rider, which would then be used to issue a traffic violation ticket. The image processing algorithm would consist of five parts, including image procurement, preliminary processing, fringe detection and segmentation, feature extraction, and recognition of character number plates using suitable machine learning algorithms. This automated system would not only make the process of issuing traffic violation tickets faster and more efficient, but it would also increase the compliance of two wheeler riders with helmet laws. This would lead to a reduction in accidents and fatalities caused by not wearing helmets, ultimately making roads safer for everyone

Analysis of Risky Riding Behavior Characteristics of the Related Road Traffic Injuries of Electric Bicycle Riders.

Electric bicycle (EB) riders, being vulnerable road users (VRUs), are increasingly becoming victims of road traffic injuries (RTIs). This study aimed to determine the current status and epidemiological characteristics of RTIs among EB riders through a questionnaire survey and roadside observations in Shantou to provide a scientific basis for the prevention and control of electric bicycle road traffic injuries (ERTIs). A total of 2412 EB riders were surveyed, and 34,554 cyclists were observed in the study. To analyze the relationship between riding habits and injuries among EB riders, chi-square tests and multi-factor logistic regression models were employed. The findings reveal that the prevalence of ERTIs in Shantou was 4.81%, and the most affected group was children under 16 years old, accounting for 9.84%. Risky behavior was widespread among EB riders, such as the infrequent wearing of safety helmets, carrying people on EBs, riding on sidewalks, and listening to music with headphones while bicycling. Notably, over 90% of those who wore headphones while bicycling engaged in this risky behavior. The logistic regression analysis showed that honking the horn (odds ratio (OR): 2.009, 95% CI: 1.245-3.240), riding in reverse (OR: 4.210, 95% CI: 2.631-6.737), and continuing to ride after a fault was detected (OR: 2.010, 95% CI: 1.188-3.402) all significantly increased the risk of ERTIs (all p < 0.05). Risky riding behavior was significantly less observed at traffic intersections with traffic officers than at those without (all p < 0.001).

AUTOMATIC E-CHALLAN GENERATING SYSTEM FOR NOT WEARING HELMET

All bike riders must now wear helmets due to the high accident rate and hazardous road conditions. Nowadays, wearing a helmet is required. Bike riders occasionally disobey safety regulations, such as not donning a helmet while riding, necessitating human intervention, which has not been very successful. Increasing the use of helmets when riding is the goal of our project. This classification- based model has undergone thorough training. Prior to identifying the number plate, the helmet is first recognized. In order to identify whether the rider is wearing a helmet, CNN recognizes the motorcycle. If the rider is not wearing a helmet, Tesseract OCR will find the motorcycle license plate. As a result, the obtained characters will be kept in the Database. Currently, classified automobiles are retrieved from the database and verified against their original identities, which are saved as details about specific vehicles. Now that the information has been matched, it is further contacted via APIs and messages are sent to the appropriate email or phone number. KEYWORDS: Machine Learning, Object detection, Neural Networks, Database.

Evaluation of Neck Pain and Posture of Cervical Spine among Helmet Users Riding Motor Bike

Background: Helmet use is become mandatory by the Indian govt. in the year of 2016. After the helmet law the prevalence of helmet users increased up to 80%. Other 20% populations refused to wear helmet while riding due their own reasons such as some founds wearing of helmet causing suffocating, heaviness over neck muscles or some feels headache too. Among various postural disorders, forward head posture (FHP) has found to be commonest. It occurs when position of the head is forward to a plumb line passing over the centre of gravity. The good posture could be present, as the external auditory meatus allied with the plumb line. FHP can be assessed by using various assessment methods and Cranio-vertebral angle (CVA) method is highly reliable method to measure FHP, which indicates smaller the angle greater the FHP. The aim of the study was to evaluate the postural changes in cervical spine and presence of neck pain among helmet users riding motor bike. Method Total 300 subjects participated in the study and were assessed by self- made questionnaire, out of which 260(Male= 189, Female= 71) subjects were taken in the study as they were fulfilling the inclusion criteria. Then dividing them into two groups based on complaining of pain on neck (with neck pain and without neck pain) using Numeric pain rating scale (NPRS). CVA was measured using MB-Ruler software for both the groups. Result The results shows CVA has a significant correlation with helmet weight (p value=0.000), BMI (p value=0.04), years of riding (p value=0.037) and days of riding per week (p value=0.031). NPRS showed to have significant correlation with helmet weight (p value=0.046) and weight of the bag (p value=0.012). 45% of study populations were having FHP and 66% populations were having neck pain. Conclusion The study concludes that helmet weight has significant impact on both neck pain and forward head posture. The incidence rate of forward head posture was greater among with neck pain populations.

Bicycle accidents in the course of time in Münster, Germany : Analysis of changes in bicycle accidents over the past 10years

Bicycles have been apopular means of transport for many years. Especially in times of increased climate discussion, the bicycle has moved further into focus as an environmentally friendly and cost-effective means of transport. Bike lanes and roads are becoming more crowded and new means of transport such as pedelecs or e‑scooters are emerging. Atotal of four studies by Münster University Hospital on bicycle accidents and the official accident statistics of the Münster police were compared. In the period from 2009 to 2019, three studies were conducted that considered all bicycle accidents and one study that only considered pedelec riders separately. The age distribution as well as main causes of accidents remained almost the same over the years. The number of pedelec accidents has increased. Pedelec riders have ahigher average age and ahigher proportion of intensive care stays; however, this cohort also has ahigh rate of helmet wearers. Overall, the willingness to wear ahelmet seems to have increased. It should be considered that with increasing bicycle traffic, safety measures must be increased accordingly. In this respect, accident prevention should focus on three major areas, engineering, education and enforcement.

Outcome of head injury in motorbike riders.

To determine the impact of helmet wearing on traumatic brain injury. We analyzed 400 cases of traumatic brain injury (TBI) in motorbike riders with and without helmet, from July 2017 to December 2020 presenting to the neurosurgery department at Jinnah Postgraduate Medical Center (JPMC), Karachi, Pakistan. The medical records were analyzed for CT scan findings, length of hospital stay, complications (mortality and disability), Glasgow Coma Scale (GCS) and Glasgow outcome score (GOS) at time of discharge. A total of 400 patients with head injury due to motorbike accidents were included and all were male patients. They were equally divided into two groups, 200 in Group-A (with helmet) and 200 in Group-B (without helmet). Majority of the unhelmeted patients i.e. 102 (51%), needed admission in the Intensive Care Unit (ICU) compared to 70 (35%) in helmeted. When comparing non-helmeted patients to helmeted patients, the total median length of hospital stay was greater among non-helmeted patients (10 vs 05 days). Mortality was higher among non-helmeted patients seen in 50 (25%) as compared to 14 (7%) in helmeted patients. Overall, the good outcome was observed in 119 (59.5%) patients in Group-A as compared to70 (35%) patients in Group-B while 81 (40.5%) showed bad outcome in Group-A and 130 (64%) in Group-B. The failure to wear a helmet was found to be strongly linked with abnormal neuroimaging more complications, poor outcome and lower GCS on discharge as compared to patients using helmet. Lack of helmet use is linked to abnormal brain imaging, more complications, and a longer stay in the hospital after a head injury.

Lightweight Helmet Detection Algorithm Using an Improved YOLOv4.

Safety helmet wearing plays a major role in protecting the safety of workers in industry and construction, so a real-time helmet wearing detection technology is very necessary. This paper proposes an improved YOLOv4 algorithm to achieve real-time and efficient safety helmet wearing detection. The improved YOLOv4 algorithm adopts a lightweight network PP-LCNet as the backbone network and uses deepwise separable convolution to decrease the model parameters. Besides, the coordinate attention mechanism module is embedded in the three output feature layers of the backbone network to enhance the feature information, and an improved feature fusion structure is designed to fuse the target information. In terms of the loss function, we use a new SIoU loss function that fuses directional information to increase detection precision. The experimental findings demonstrate that the improved YOLOv4 algorithm achieves an accuracy of 92.98%, a model size of 41.88 M, and a detection speed of 43.23 pictures/s. Compared with the original YOLOv4, the accuracy increases by 0.52%, the model size decreases by about 83%, and the detection speed increases by 88%. Compared with other existing methods, it performs better in terms of precision and speed.

Parental beliefs regarding adult and child motorcycle helmet-wearing practices in Vietnam.

From a global perspective, low helmet-wearing rates in children are an ongoing concern. International studies show adult helmet-wearing rates are consistently higher than those of children. Research also shows parents are highly influential in promoting healthy behavior in their children. Parent-focused helmet-wearing campaigns are a practical way to actively feature parents influencing helmet-wearing in children. The current study obtained anonymous parent-focused survey data on parent and child helmet-wearing practices, perceived affects of helmet use on children's health, and beliefs about brain injury prevention from a sample of Vietnams parents. Findings from this study are consistent with previous international research that shows adults wear helmets at higher rates than children. Data from this study may help those responsible for child helmet safety campaigns develop more effective child helmet-wearing campaigns, particularly in Southeast Asian countries.

Epidemiological characteristics of crashes and pattern of motorcycle injuries presenting to hospitals in Kisumu City, Kenya: a cross-sectional study.

motorcycle injuries comprise considerable morbidity, disability and mortality of road traffic casualties. The study aimed to assess the pattern and characteristics of motorcycle crash injuries that sought care at the Emergency Departments in Tier III hospitals in Kisumu City during a six-month period from May to November 2019. this was a cross-sectional study of all motorcycle injury patients presenting to three tier III public and private hospitals in Kisumu City. Using a structured questionnaire, data were collected on human and injury characteristics. Multiple logistic regression model was used to determine the predictors of fatality. Statistical significance was set at p<0.05. a total of 1073 of motorcycle crash injury patients sought and obtained care at the hospitals. The majority (73.6%) were males. Seventy-three cases resulted in death (case-fatality rate of 6.80%. The age range was 2-84 years, with a mean of 29.6 years (± SD 12.19). Majority, (43.3%) were aged 21-30 years. Most of the crashes occurred during the daytime (79.1%). Of all motorcycle riders and pillion passengers 30.6% wore helmets at the time of the crash. Head injuries (43.6%) were the most common. Injury seventy scores (ISS) ranged from 1 to 51. Glasgow comma scale <3, un-helmeted patients and major trauma (ISS > 16), significantly influenced mortality (P< 0.001). these findings confirmed and strengthened the documented substantial morbidity, mortality that motorcycle crashes place on individuals and society, contribute to the body of literature on motorcycle injuries and potentially assist in policy decisions on motorcycle transport safety.

A Study of Injury Pattern and Mode of Accident of Fatal Injuries in RTA Cases Admitted in a Tertiary Care Centre -3 Year Retrospective Study

Road traffic accident ranks among the top causes of death in the world; after ischemic heart disease, it is projected to become the second leading cause in 2020. A report states that 1.24 million people die every year worldwide on the roads. RTA is a public health issue that greatly affects individuals, families, communities, and nations. The cost of burden is estimated to be around 1–2% of a country’s GNPin low-income countries. To know the pattern of fatal injuries in RTA cases. To correlate the survival period and cause of death in fatal road traffic accidents. Data from medico-legal autopsies of all RTA victims from 2018 to 2020 (n= 210) were collected in this retrospective study. Data were analyzed using SPSS software, V.22. The majority of the study population, 24.2% (51 out of 210), belonged to the 1-30 age group. Male predominance was noted in the study with 84.2% (177 out of 210). The type of accident was self-fall in high proportion 26.6% (56 out of 210) followed by a collision between 2-wheelers and 4-wheelers 24.2% (51 out of 210). Intracranial haemorrhage was reported in 140 (66.6%) study participants as the leading cause of death followed by hemorrhagic shock other than head and neck injury reported in 42 (19.9%) cases. A short revival period of fewer than 24 hours was noticed in intra-cranial haemorrhage (81 out of 140) followed by hemorrhagic shock other than head and neck injury (39 cases). Skull fractures are higher in 2-wheelers compared to other types of vehicles and pedestrians. (103 out of 140). The majority of fatal accidents occurred in two-wheelers. Hence, health education and awareness should be created among two-wheeler riders, especially drivers &amp; pillion to wear helmets and follow safety measures while riding.

Determinants of switching behavior to wear helmets when riding e-bikes, a two-step SEM-ANFIS approach.

E-bikes have become one of China's most popular travel modes. The authorities have issued helmet-wearing regulations to increase wearing rates to protect e-bike riders' safety, but the effect is unsatisfactory. To reveal the factors influencing the helmet-wearing behavior of e-bike riders, this study constructed a theoretical Push-Pull-Mooring (PPM) model to analyze the factor's relationship from the perspective of travel behavior switching. A two-step SEM-ANFIS method is proposed to test relationships, rank importance and analyze the combined effect of psychological variables. The Partial Least Squares Structural Equation Model (PLS-SEM) was used to obtain the significant influencing factors. The Adaptive Network-based Fuzzy Inference System (ANFIS), a nonlinear approach, was applied to analyze the importance of the significant influencing factors and draw refined conclusions and suggestions from the analysis of the combined effects. The PPM model we constructed has a good model fit and high model predictive validity (GOF = 0.381, R2 = 0.442). We found that three significant factors tested by PLS-SEM, perceived legal norms (β = 0.234, p < 0.001), perceived inconvenience (β = -0.117, p < 0.001) and conformity tendency (β = 0.241, p < 0.05), are the most important factors in the effects of push, mooring and pull. The results also demonstrated that legal norm is the most important factor but has less effect on people with low perceived vulnerability, and low subjective norms will make people with high conformity tendency to follow the crowd blindly. This study could contribute to developing refined interventions to improve the helmet-wearing rate effectively.

Helmet use in skiing and snowboarding in Switzerland during the past 20 years

The aim of this study was to quantify helmet use in skiing and snowboarding in Switzerland. Regular cross-sectional surveys of skiers and snowboarders were conducted on Swiss slopes from the 2002/2003 to the 2021/2022 seasons. Surveys included information on helmet use, sport, gender, age, skill level, and country of residence. Data from 15 single surveys were used to trace the evolution of helmet use in Switzerland. A logistic regression on data from the most recent survey was conducted to detect differences in helmet use among demographic groups. In the 2002/2003 season, 14% of skiers (95% confidence interval [CI] ​= ​12%–18%) and 20% of snowboarders (CI 16%–26%) used a helmet. By the most recent season (2021/2022), wearing rates for both skiers and snowboarders had increased significantly (skiers: 95% [CI 93%–97%]; snowboarders: 92% [CI 89%–95%]). The regression results showed that skiers, females, younger people, and people in the German-speaking part of Switzerland were more likely to wear a helmet. The results show an impressive evolution of helmet use rates among skiers and snowboarders since the 2002/2003 season. Steep differences detected in earlier seasons have leveled out, and most people on the slopes today wear helmets. Nevertheless, the ongoing efforts for promoting helmet use should be maintained.

Smart Helmet

Abstract: As the number of motorcyclists in our country rises, so do the number of traffic accidents and fatalities. The majority of these accidents are the result of drunk driving and a failure to wear a helmet. The majority of nations now require their residents to wear helmets while riding bikes and to never ride while intoxicated, but despite this, the laws are still broken. Drive Protection and Accident Detection Smart Helmet was developed in an effort to address this issue as engineers using the use of mechatronics. It is made up of an intelligent system that is built into the helmet and the vehicle. The helmet device makes sure the motorcyclist is wearing a helmet and is not drinking during the trip. If the aforementioned requirement is not met, it interacts with the vehicle unit to turn off the motorcycle's ignition system. Vehicle unit checks and intimates’ accident and SMS notifies of accident using geometric coordinates. Geometric coordinates can be used to locate the injured cyclist using a basic GPS monitoring software. The primary goal of the suggested idea is to offer a safe and affordable smart helmet. The smart helmet was designed using a Wi-Fi enabled processor and an integrated network of sensors for the engine control system, accident alert system, and alcohol detection. When it comes to the driver's safety, the suggested system is quite beneficial.

Risk-taking behaviors of the Tehran city; Iranian college students in 2018

Background and aims: High-risk behaviors among different strata are one of the most severe health threats in recent years. This study aimed to investigate the frequency of high-risk behaviors among college students in Iran. Methods: This study was a cross-sectional study with 144 new students in Tehran City, Iran, performed in 2018 using random sampling. The data collection tools included questions on demographic variables and Youth Risk Behavior Survey (YRBS) questionnaire. Students completed the questionnaire. The data were analyzed using SPSS 24 and descriptive tests. Results: In this study, all students were in their first year of university. The results showed that 66% of students were girls and 34% were boys. 6.9% and 9% of students did not wear helmets when riding motorcycles or sitting in the driver’s seat, respectively. Also, threatened and beaten once with a weapon such as a knife or a stick and physically assaulted or beaten once were seen in 9% and 3.6% of students, respectively.9.7% had taken money from their parents or others once without permission. Conclusion: Based on the findings, many participants engage in high-risk behaviors that endanger their health; Therefore, designing and improving health programs and strategies is essential to reduce the risks and factors that cause high-risk behaviors.

Pediatric injury due to wheeled recreational devices: a single-institution retrospective study

BackgroundInjuries are the number one cause of death in children and cause significant morbidity. Common scenarios for injury include wheeled recreational devices (WRDs) that allow children to be mobile and independent (example ATV-all terrain vehicles, dirt bikes, bicycles, skateboards, and scooters). We present a case series review of these external causes of injury. This study aims to evaluate epidemiologic trends in WRD injuries and patterns in usage of protective gear.ResultsA total of 263 patients were identified as meeting criteria for inclusion with the following causes of injuries-103 bicycle, 73 ATV, 27 dirt bike, 14 skateboard, 13 motorcycle, 7 go carts, 3 hover board, 3 roller skates, 1 dune buggy, 1 motor scooter, 1 rip stick, and 1 tractor toy. Ages of patients ranged from 2 to 18 years of age with the greatest range being noted for bicycles (2–17 years) and motorcycles (3–18 years). The mean age was higher for motorcycle and skateboard (12.9 and 11.6, respectively) and lowest for scooter and bicycle (8.3 and 9.2, respectively). The majority of [overall study (97%) and for each mode of transportation] patients were Caucasian, which is in contrast to our overall ED population, which is only 42% Caucasian. The majority of patients were male [190 (72%)]. Over half of the overall patients, 159 of the 263 (60%), were not wearing a helmet (with only 10 charts having no documentation of helmet use). In regard to ATV riders specifically, 58% were not wearing helmets at the time of injury, with an additional 5 patients who reported their helmet came off. The lowest percentage of riders reporting appropriate helmet use was skateboarders with only 21% wearing helmets, and the highest percentage was dirt bike riders with 74% reporting helmet use.ConclusionCommon scenarios for injury include WRDs that allow children to be mobile and independent. Most of these injuries were found in Caucasian males between the ages of 9–12 with low rates of helmet use. This study adds to the literature with a description of the breadth of products children use and are injured while using.

All-terrain vehicle exposure and the association of certified training on adolescent safety behaviors and crash experiences

BackgroundCertified training courses in all-terrain vehicle (ATV) operation are recommended, but little has been published regarding how they affect riding behaviors. Our objectives were to determine adolescents’ ATV riding exposures, crashes and injuries, and the association of completing certified ATV training on riding behaviors.MethodsRespondents completed an anonymous paper survey at 30 school districts approximately one year after participating in the Safety Tips for ATV Riders (STARs) program. Survey data were compiled using Qualtrics™. Frequency, chi-square, Fisher’s exact test, and logistic regression analyses were performed using SAS, V.9.4.Results4,891 students completed the survey from Fall 2012–Fall 2019. Respondents were 10–18 years old, and similar numbers of participants were male and female. Fifty-nine percent lived in town, 18% on a farm, and 22% in the country but not on farm. Forty-two percent reported their families owned an ATV. Seventy-seven percent had ridden an ATV, 40% of whom reported riding at least weekly. The vast majority (94%) used ATVs for recreation and 49% used them for work purposes. In the previous year, 22% of riders reported having been in a crash, with 7% of crash victims requiring medical attention for injuries. Greater crash likelihood was seen in males, recreational riders, more frequent riders, competitive racers and those who had ridden with passengers or on roads. Only 8% of riders had completed certified ATV safety training (note, STARs is not a certification program). Those whose families owned ATVs, more frequent riders, and public park users had greater likelihoods of course completion. Relative to their peers, respondents who completed a certification course had higher proportions that always/almost always wore helmets (39% vs. 20%, p < 0.0001) and lower proportions that never/almost never wore helmets (29% vs 58%, p < 0.0001), had ridden with passengers (63% vs. 96%, p < 0.0001), and had driven on public roads (41% vs 50%, p = 0.0065).ConclusionsATV safety training certification among Iowa adolescents in the study was infrequent but those that received training reported higher helmet use, less riding with passengers, and less driving on public roads. These data suggest completing safety training certification may promote safer riding behaviors among youth.

Investigation into Recognition Technology of Helmet Wearing Based on HBSYOLOX-s

This work proposes a new approach based on YOLOX model enhancement for the helmet-wearing real-time detection task, which is plagued by low detection accuracy, incorrect detection, and missing detection. First, in the backbone network, recursive gated convolution (gnConv) is utilized instead of traditional convolution, hence addressing the issue of extracting many worthless features due to excessive redundancy in the process of feature extraction using conventional convolution. Replace the original FPN layer in the Neck network with the EfficientNet-BiFPN layer. Realize top–down and bottom–up bidirectional fusion of deep and shallow features to improve the flow of feature data between network layers. Lastly, the SIOU cross-entropy loss function is implemented to address the issue of missed detections in crowded environments and further increase the model’s detection precision. Experiments and data comparisons indicate that the modified model’s average detection accuracy is 95.5%, which is 5.4% higher than that of the original network model, and that the detection speed has been dramatically increased to suit actual production requirements.