Machine Vision Technology Research Articles

Weed Management Methods for Herbaceous Field Crops: A Review

Weeds compete with crops for water and nutrients and can adversely affect crop growth and yield, so it is important to research effective weed control methods. This paper provides an overview of the impact of weeds on crop yield and describes the current state of research on weed management in field herbaceous crops. Physical weed control mainly refers to thermal technologies represented by flame weed control and laser weed control, which can efficiently and accurately remove weeds. Mechanical weed control requires a combination of sensor technologies, machine vision technology, and high-precision navigation to improve weed control accuracy. Biological weed control relies heavily on plant extracts and pathogens to create herbicides, but it is costly, and some can be toxic to mammals. Chemical weed control is a common method, resulting in environmental pollution and weed resistance. To reduce the use of chemical herbicides, scholars have proposed integrated weed management strategies, which combine biological control, control of the seed bank, and improve crop competitiveness. Integrated weed management strategies are considered to be the future direction of weed management. In conclusion, physical, mechanical, biological, and chemical weed control methods are commonly used in weed management. Each method has its applicable scenarios, and the implementation of integrated weed management strategies can lead to better weed control, improving crop yield and quality. The main objective of this review is to organize the research progress on weed management methods for herbaceous crops in the field and to provide a reference for the agricultural sector to develop weed control strategies. Specifically, this paper categorizes weed management methods into four groups, discusses and presents the advantages and disadvantages of the aforementioned weed control methods, and discusses future research directions.

Pedicle screw accuracy placed with assistance of machine vision technology in patients with neuromuscular scoliosis.

Pedicle screws are the primary method of vertebral fixation in scoliosis surgery, but there are lingering concerns over potential malposition. The rates of pedicle screw malposition in pediatric spine surgery vary from 10% to 21%. Malpositioned screws can lead to potentially catastrophic neurological, vascular, and visceral complications. Pedicle screw positioning in patients with neuromuscular scoliosis is challenging due to a combination of large curves, complex pelvic anatomy, and osteopenia. This study aimed to determine the rate of pedicle screw malposition, associated complications, and subsequent revision from screws placed with the assistance of machine vision navigation technology in patients with neuromuscular scoliosis undergoing posterior instrumentation and fusion. A retrospective analysis of the records of patients with neuromuscular scoliosis who underwent thoracolumbar pedicle screw insertion with the assistance of machine-vision image guidance navigation was performed. Screws were inserted by either a staff surgeon, orthopaedic fellow, or orthopaedic resident. Post-operative ultra-low dose CT scans were used to assess pedicle screw accuracy. The Gertzbein classification was used to grade any pedicle breaches (grade 0, no breach; grade 1, <2 mm; grade 2, 2-4 mm; grade 3, >4 mm). A screw was deemed accurate if no breach was identified (grade 0). 25 patients were included in the analysis, with a mean age of 13.6 years (range 11 to 18 years; 13/25 (52.0%) were female. The average pre-operative supine Cobb angle was 90.0 degrees (48-120 degrees). A total of 687 screws from 25 patients were analyzed (402 thoracic, 241 lumbosacral, 44 S2 alar-iliac (S2AI) screws). Surgical trainees (fellows and orthopaedic residents) inserted 46.6% (320/687) of screws with 98.8% (4/320) accuracy. The overall accuracy of pedicle screw insertion was 98.0% (Grade 0, no breach). All 13 breaches that occurred in the thoracic and lumbar screws were Grade 1. Of the 44 S2AI screws placed, one screw had a Grade 3 breach (2.3%) noted on intra-operative radiographs following rod placement and correction. This screw was subsequently revised. None of the breaches resulted in neuromonitoring changes, vessel, or visceral injuries. Machine vision navigation technology combined with careful free-hand pedicle screw insertion techniques demonstrated high levels of pedicle screw insertion accuracy, even in patients with challenging anatomy.

Measurement Method of Bar Unmanned Warehouse Area Based on Binocular Vision

With the development of Industry 4.0 and the implementation of the 14th Five-Year Plan, intelligent manufacturing has become a significant trend in the steel industry, which can propel the steel industry toward a more intelligent, efficient, and sustainable direction. At present, the operation mode of unmanned warehouse area for slabs and coils has become relatively mature, while the positioning accuracy requirement of bars is getting more stringent because they are stacked in the warehouse area according to the stacking position and transferred by disk crane. Meanwhile, the traditional laser ranging and line scanning method cannot meet the demand for precise positioning of the whole bundle of bars. To deal with the problems above, this paper applies machine vision technology to the unmanned warehouse area of bars, proposing a binocular vision-based measurement method. On the one hand, a 3D reconstruction model with sub-pixel interpolation is established to improve the accuracy of 3D reconstruction in the warehouse area. On the other hand, a feature point matching algorithm based on motion trend constraint is established by means of multi-sensor data fusion, thus improving the accuracy of feature point matching. Finally, a high-precision unmanned 3D reconstruction of the bar stock area is completed.

Anterior pelvic tilt increases hamstring strain and is a key factor to target for injury prevention and rehabilitation.

Hamstring muscle strain injury is very common in sports involving high-speed running. Hamstring muscles originate from the ischial tuberosity and thus pelvic position may influence hamstring strain during different sports movements like sprinting, but this has only been evaluated by indirect methods. This study tested the hypothesis that a change in anterior pelvic tilt causes elongation of the overall hamstring complex and disproportionately elongates proximal relative to distal muscle regions. Seven fresh-frozen specimens (full lower limb with pelvis and lumbar spine) were used for this in vitro study. Specimens were dissected to enable visualization of the hamstring muscles and then fixed into a custom-made testing bench that allowed controlled movement of the pelvis over a fixed femur and tibia. Nine markers were inserted into the hamstring muscles to allow intra- and intermuscle difference measurements. Then, six different anterior pelvic angles were used to measure the difference in hamstring muscle lengthening through a three-dimensional reconstruction system based on stereoscopic machine vision technology. An increase in anterior pelvic tilt produced a significant non-uniform increase in tissue elongation in all regions of the three hamstring muscles (semitendinosus, semimembranosus [SMB] and biceps femoris long head), which was greater in the proximal (>1 cm every 5°) compared to the distal region (≈0.4 cm every 5°). At the proximal hamstring region, SMB showed significantly greater length changes compared to conjoint tendons with nonstatistically significant elongation differences between muscles at the distal region. Considering the results of the study, the pelvis segment will likely play a fundamental role as a strain regulator of hamstring muscles. These results will have an impact on injury rehabilitation and prevention processes of hamstring injuries, as well as optimize future musculoskeletal models and avoid potential underestimation of the hamstring muscle-tendon complex lengthening during high-speed running. N/A.

Maturity detection of single maize seeds based on hyperspectral imaging and transfer learning

Seed maturity is closely related to seed germination, seed storage and the yield of the next generation of crops and it is an important indicator of seed quality. Maize is an important food, feed, and industrial raw material crop in China. Hyperspectral imaging technology has the advantages of spectral technology and machine vision technology and it can obtain information inside and outside the seed without damaging it. On the basis of hyperspectral imaging technology, transfer learning algorithms were used in this study to explore the transfer feasibility of maize seed maturity models amongst different varieties. Three transfer tasks were designed between different varieties on the basis of the model established for the Xianyu 335 variety. Cross-variety detection models were constructed using the classical machine-learning algorithm partial least square-linear discriminant analysis (PLS-LDA), the transfer learning algorithm transfer component analysis (TCA), and manifold embedded distribution alignment (MEDA). Results showed that the PLS-DA direct transfer model had varying degrees of model failure in all three transfer tasks. However, the average recognition accuracies of the TCA migration model and the MEDA migration model were 89.3% and 91.2%, respectively. Both of the two Transfer learning models can effectively solve the problem of different data feature distributions amongst different varieties, making the model more universal. It solves the problem of transferring the single-grain maize seed maturity detection model amongst different varieties, providing effective, stable and universal technical support for maize seed detection, which is conducive to the promotion of single-grain sowing technology and the improvement of agricultural automation level.

Automatic fruit picking technology: a comprehensive review of research advances

In recent years, the fruit industry has become an important part of agricultural development, and fruit harvesting is a key stage in the production process. However, picking fruits during the harvest season is always a major challenge. In order to solve the challenges of time-consuming, costly, and inefficient fruit picking, researchers have conducted a lot of studies on automatic fruit picking equipment. Existing picking technologies still require further research and development to improve efficiency and reduce fruit damage. Aiming at the efficient and non-destructive picking of fruits, this paper reviews machine vision and mechanical fruit picking technology and the current research status, including the current application status, equipment structure, working principle, picking process, and experimental results. As a promising tool, machine vision technology has been widely researched and applied due to its low hardware cost and rich visual information. With the development of science and technology, automated fruit picking technology integrates information technology, integrates automatic perception, transmission, control, and operation, etc., saves manpower costs, and continuously promotes the development of modern agriculture in the direction of refinement of equipment technology, automation, and intelligence. Finally, the challenges faced by automated fruit picking are discussed, and future development is looked forward to with a view to contributing to its sustainable development.

Hardware and Software Solutions for the Generation of a Database of HSV-Color Characteristics for the Main Ores and Rocks of the Khibiny Massif

This article presents developed hardware and software solutions based on the application of machine vision technology. The hardware and software solutions were created in order to generate a database of HSV-color value for the main ore types, host rocks, and minerals to define criteria for the in-process identification of the Khibiny apatite in testing the walls of blasting boreholes. The hardware ensures a multi-parametric assessment of the optical characteristics of samples and minerals located on their surface. The designed software solution allows the user to control the measurement process; systematize a description of the textural and structural features of the sample under study; and process images of the core surface. The resultant database of HSV-color value for the main ores and rocks of the Khibiny massif and their constituent minerals will provide an opportunity to search for criteria for the in-process identification of the Khibiny apatite in a mineral mixture.

Revolutionizing Precision Agriculture: A Comprehensive Review of Innovative Technologies and Application in Digital Farming

Precision agriculture has witnessed a revolution propelled by innovative technologies and applications in digital farming. This comprehensive review explores the transformative impact of these advancements on agricultural practices. Cutting-edge technologies such as agricultural robotics, mechatronic platforms, and autonomous systems are examined for their role in enhancing efficiency and productivity in farming operations. The integration of low-cost location sensing subsystems and adaptive systems further optimizes precision agriculture, enabling real-time decision-making and resource management. Deep learning techniques, specifically in plant disease detection and leaf segmentation from digitized herbarium specimen images, demonstrate remarkable potential for disease management and crop monitoring. Additionally, the design and simulation of robotic arms using machine vision technology represent significant strides in automating tasks within agricultural settings. Through a critical analysis of these technologies and their applications, this review underscores the transformative power of digital farming in revolutionizing precision agriculture. It highlights the potential to address key challenges in sustainable agriculture, improve resource utilization, and meet the growing demands for food production in a rapidly evolving agricultural landscape.

Research Progress on Quality Detection of Livestock and Poultry Meat Based on Machine Vision, Hyperspectral and Multi-Source Information Fusion Technologies

Presently, the traditional methods employed for detecting livestock and poultry meat predominantly involve sensory evaluation conducted by humans, chemical index detection, and microbial detection. While these methods demonstrate commendable accuracy in detection, their application becomes more challenging when applied to large-scale production by enterprises. Compared with traditional detection methods, machine vision and hyperspectral technology can realize real-time online detection of large throughput because of their advantages of high efficiency, accuracy, and non-contact measurement, so they have been widely concerned by researchers. Based on this, in order to further enhance the accuracy of online quality detection for livestock and poultry meat, this article presents a comprehensive overview of methods based on machine vision, hyperspectral, and multi-sensor information fusion technologies. This review encompasses an examination of the current research status and the latest advancements in these methodologies while also deliberating on potential future development trends. The ultimate objective is to provide pertinent information and serve as a valuable research resource for the non-destructive online quality detection of livestock and poultry meat.

Vehicle Load Identification Using Machine Vision and Displacement Influence Lines

In recent years, bridge collapses resulting from vehicle overloading have underscored the crucial necessity for real-time monitoring of traffic conditions on bridges, making pavement-based weigh-in-motion systems indispensable for large bridges. However, these systems usually have poor durability and will cause traffic interruptions during their installation and maintenance processes. This paper addresses the challenge of recognizing vehicle loads by proposing a vehicle load identification method based on machine vision and displacement influence lines. The technology consists of three essential steps. Firstly, machine vision technology is utilized to identify vehicle trajectories. Following this, the displacement response, monitored by millimeter-wave radar, is integrated to calculate the influence lines of the structure’s displacement. Lastly, an overall least squares method incorporating a regularization term is applied to calculate axle weights. The efficacy of the proposed method is validated within the monitoring system of a specific continuous beam. Importantly, the calibration of vehicles and the validation dataset rely on information monitored by the pavement-based weigh-in-motion system of adjacent arch bridges, serving as ground truth. Results indicate that the identification errors for gross vehicle weight do not exceed 25%. This technology holds significant importance for identifying vehicle weights on small to medium-span bridges. Due to its cost-effectiveness, easy installation, and maintenance, it possesses a high potential for widespread adoption.

Lightweight Pig Face Feature Learning Evaluation and Application Based on Attention Mechanism and Two-Stage Transfer Learning

With the advancement of machine vision technology, pig face recognition has garnered significant attention as a key component in the establishment of precision breeding models. In order to explore non-contact individual pig recognition, this study proposes a lightweight pig face feature learning method based on attention mechanism and two-stage transfer learning. Using a combined approach of online and offline data augmentation, both the self-collected dataset from Shanxi Agricultural University's grazing station and public datasets underwent enhancements in terms of quantity and quality. YOLOv8 was employed for feature extraction and fusion of pig face images. The Coordinate Attention (CA) module was integrated into the YOLOv8 model to enhance the extraction of critical pig face features. Fine-tuning of the feature network was conducted to establish a pig face feature learning model based on two-stage transfer learning. The YOLOv8 model achieved a mean average precision (mAP) of 97.73% for pig face feature learning, surpassing lightweight models such as EfficientDet, SDD, YOLOv5, YOLOv7-tiny, and swin_transformer by 0.32, 1.23, 1.56, 0.43 and 0.14 percentage points, respectively. The YOLOv8-CA model’s mAP reached 98.03%, a 0.3 percentage point improvement from before its addition. Furthermore, the mAP of the two-stage transfer learning-based pig face feature learning model was 95.73%, exceeding the backbone network and pre-trained weight models by 10.92 and 3.13 percentage points, respectively. The lightweight pig face feature learning method, based on attention mechanism and two-stage transfer learning, effectively captures unique pig features. This approach serves as a valuable reference for achieving non-contact individual pig recognition in precision breeding.

Digital Model of Plan View Pattern Control for Plate Mills Based on Machine Vision and the DBO-RBF Algorithm

Plan view pattern control (PVPC) is a highly effective means to improve the rectangularization of products and increase the yield of plate mills. By optimizing the parameters of PVPC, the effect of PVPC can be further improved. In this paper, a digital model for predicting and controlling crop patterns of plates is proposed based on the radial basis function (RBF) neural network optimized by the dung beetle optimizer (DBO) algorithm. Machine vision technology is used to obtain a digital description of the crop pattern of the rolled plates. An automatic threshold adjustment algorithm is proposed for the image processing of plate pattern photos during the rolling process. The error between the pattern data calculated through machine vision technology and the measured pattern data does not exceed 3 mm. The spread parameters of the RBF are optimized using DBO, and the digital model structure is established. The goodness of fit (R2) and the mean absolute error (MAE) are used as evaluation indicators. The results show that the digital model established based on DBO-RBF has good predictive and control performance, realizing intelligent prediction of the crop pattern of plates and the parameter optimization of PVPC. In practical applications, the crop cutting loss area of irregular deformation at the end of the plate can be reduced by 31%.

Multichannel meta-imagers for accelerating machine vision.

Rapid developments in machine vision technology have impacted a variety of applications, such as medical devices and autonomous driving systems. These achievements, however, typically necessitate digital neural networks with the downside of heavy computational requirements and consequent high energy consumption. As a result, real-time decision-making is hindered when computational resources are not readily accessible. Here we report a meta-imager designed to work together with a digital back end to offload computationally expensive convolution operations into high-speed, low-power optics. In this architecture, metasurfaces enable both angle and polarization multiplexing to create multiple information channels that perform positively and negatively valued convolution operations in a single shot. We use our meta-imager for object classification, achieving 98.6% accuracy in handwritten digits and 88.8% accuracy in fashion images. Owing to its compactness, high speed and low power consumption, our approach could find a wide range of applications in artificial intelligence and machine vision applications.

Research on an Online Monitoring Device for the Powder Laying Process of Laser Powder Bed Fusion.

Improving the quality of metal additive manufacturing parts requires online monitoring of the powder bed laying procedure during laser powder bed fusion production. In this article, a visual online monitoring tool for flaws in the powder laying process is examined, and machine vision technology is applied to LPBF manufacture. A multiscale improvement and model channel pruning optimization method based on convolutional neural networks is proposed, which makes up for the deficiencies of the defect recognition method of small-scale powder laying, reduces the redundant parameters of the model, and enhances the processing speed of the model under the premise of guaranteeing the accuracy of the model. Finally, we developed an LPBF manufacturing process laying powder defect recognition algorithm. Test experiments show the performance of the method: the minimum size of the detected defects is 0.54 mm, the accuracy rate of the feedback results is 98.63%, and the single-layer laying powder detection time is 3.516 s, which can realize the effective detection and control of common laying powder defects in the additive manufacturing process, avoids the breakage of the scraper, and ensures the safe operation of the LPBF equipment.

Steel Strip Surface Defect Detection Method Based on Improved YOLOv5s.

Steel strip is an important raw material for the engineering, automotive, shipbuilding, and aerospace industries. However, during the production process, the surface of the steel strip is prone to cracks, pitting, and other defects that affect its appearance and performance. It is important to use machine vision technology to detect defects on the surface of a steel strip in order to improve its quality. To address the difficulties in classifying the fine-grained features of strip steel surface images and to improve the defect detection rate, we propose an improved YOLOv5s model called YOLOv5s-FPD (Fine Particle Detection). The SPPF-A (Spatial Pyramid Pooling Fast-Advance) module was constructed to adjust the spatial pyramid structure, and the ASFF (Adaptively Spatial Feature Fusion) and CARAFE (Content-Aware ReAssembly of FEatures) modules were introduced to improve the feature extraction and fusion capabilities of strip images. The CSBL (Convolutional Separable Bottleneck) module was also constructed, and the DCNv2 (Deformable ConvNets v2) module was introduced to improve the model's lightweight properties. The CBAM (Convolutional Block Attention Module) attention module is used to extract key and important information, further improving the model's feature extraction capability. Experimental results on the NEU_DET (NEU surface defect database) dataset show that YOLOv5s-FPD improves the mAP50 accuracy by 2.6% before data enhancement and 1.8% after SSIE (steel strip image enhancement) data enhancement, compared to the YOLOv5s prototype. It also improves the detection accuracy of all six defects in the dataset. Experimental results on the VOC2007 public dataset demonstrate that YOLOv5s-FPD improves the mAP50 accuracy by 4.6% before data enhancement, compared to the YOLOv5s prototype. Overall, these results confirm the validity and usefulness of the proposed model.

Design and optimization of an octuple-electrode array for micro-particle chain rotation via electrorotation integrated with machine vision technology.

Microparticle rotation is an important process in biomedical engineering, such as biosensors, cell injection or cell morphology. Single particle rotation has been widely investigated, while rotation of particle chains has gained rare attention. In this paper, we utilize a noncontact manipulation method to rotate microparticle chains via electrorotation by designing an octuple-electrode array (OEA). Finite element simulations were conducted for analyzing the desired electrode field and optimizing the structure of microelectrode pairs. The direction of the electric field in the workspace is investigated with different voltage signal inputs through specially designed circuits. In the experiment, microparticles are driven to form several chains in the proposed electrode fields. With the rotation of the electric field, particle chains could be rotated synchronously. Automated rotation and detection of polystyrene microspheres and yeast cell chains are achieved using machine vision technology. Results show that the proposed method could be utilized to rotate ordered microparticles with an appropriate input signal.

Advanced design and Engi-economical evaluation of an automatic sugarcane seed cutting machine based RGB color sensor.

There are many problems related to the use of machine learning and machine vision technology on a commercial scale for cutting sugarcane seeds. These obstacles are related to complex systems and the way the farmers operate them, the possibility of damage to the buds during the cleaning process, and the high cost of such technology. In order to address these issues, a set of RGB color sensors was used to develop an automated sugarcane seed cutting machine (ASSCM) capable of identifying the buds that had been manually marked with a unique color and then cutting them mechanically, and the sugarcane seed exit chute was provided with a sugarcane seed monitoring unit. The machine's performance was evaluated by measuring the damage index at sugarcane stalk diameters of 2.03, 2.72, 3.42, and 3.94 cm. where two different types of rotary saw knives had the same diameter of 7.0 in/180 mm the two knives had 30 and 80 teeth, also we used five cutting times of 1000, 1500, 2000, 2500, and 3000 ms. All tests were done at a fixed cutting speed of 12000 rpm. In addition, the machine's performance was evaluated by conducting an economic analysis. The obtained results showed that the most damage index values were less than 0.00 for all cutting times and sugarcane stalk diameters under testing, while the DI values were equal zero (partial damage) for sugarcane stalk diameter of 3.42 cm at cutting times of 2000 ms and 2500 ms, in addition to the DI values being equal zero (extreme damage) for sugarcane stalk diameter of 3.94 cm at cutting times of 1500 ms and 2000 ms. The economic analysis showed that the total cost of sugarcane seeds per hectare is 70.865 USD. In addition, the ASSCM can pay for itself in a short period of time. The payback time is 0.536 years, which means that the ASSCM will save enough money to pay for itself in about 6.43 months. Finally, we suggest using a rotary saw knife with 80 teeth and a cutting time of 2000 ms to cut sugarcane stacks with an average diameter of 2.72 cm. This will result in higher performance and lower operating costs for the ASSCM.

Machine Vision for Efficient Electric Vehicle Charging Station Deployment

This research examines the optimization of the deployment of electric vehicle (EV) charging stations by using machine vision technology, which involves analyzing real-time data and geographical information. Geospatial data analysis reveals prospective sites for charging stations by considering population density and accessibility to roads, hence identifying regions with increased demand for electric vehicle charging. The assessment of electric vehicle (EV) traffic patterns highlights the ever-changing charging requirements at various times and places, underscoring the need of flexible deployment techniques. Furthermore, evaluating the costs of implementing the deployment and the capabilities of charging, it becomes apparent that there are compromises to be made between the initial expenditures of installation, the amount of power generated, and the quantity of charging stations. These trade-offs are essential for optimizing resources. The usage study of charging stations using machine vision reveals variations in the number of available charging points at different stations and the need for adaptive resource distribution timestamps, techniques. The examination of percentage change reveals notable fluctuations in population density, installation costs, and the availability of charging points. This information is crucial for making well-informed decisions about the deployment of charging infrastructure. Combining machine vision insights with geographical and traffic analyses presents a promising method to create data-driven strategies for the placement of EV charging stations. This approach addresses the changing needs of electric mobility and provides guidance to stakeholders for efficient and flexible charging solutions.

Optimizing efficiency and zero-defect manufacturing with in-process inspection: challenges, benefits, and aerospace application

In this paper, we present a comprehensive study on the implementation of machine vision-enabled in-process quality inspection systems in machining operations. Our objective is to enable zero-defect manufacturing by maximizing efficiency and effectiveness in production processes. We utilize machine vision technology as a transformative tool, playing a vital role in our comparative analysis which reveals its superiority over traditional in-situ approaches in both ideal and challenging real-world scenarios. Through simulation, we demonstrate how machine vision improves the performance of in-line process systems. We also discuss substantial challenges of implementation, such as managing environmental contamination, optimizing machine coordination, accommodating a range of part sizes, and configuring effective coolant delivery systems. Our in-depth analysis of essential factors includes the robustness of machine vision equipment, operator training for machine vision technology, and cost-benefit analysis of its implementation. The research emphasizes machine vision's crucial role in transforming manufacturing setups and enhancing advanced automation systems. The study underscores the immense potential of machine vision in in-process quality inspection to significantly reduce production costs and time, fostering higher manufacturing sustainability and competitive advantage in the era of Industry 4.0.

Detection and Application of Concrete Compressive Strength with Machine Vision Technology

Abstract This paper proposes research on concrete compressive strength detection based on machine vision technology in response to the problems of traditional concrete compressive strength detection methods. The original image is directly input into the network for training. There will be too much memory occupation. The efficiency is too low, it needs to be visually preprocessed to meet the requirements of the dataset to establish a high-performance concrete compressive strength detection network model. Then, the processed high-performance concrete compressive strength feature images are inputted into the Mask R-CNN network. The compressive strength features are extracted and trained by the convolutional neural network. Then, the extracted features are further processed with a bilinear difference to ensure their integrity, and finally, the construction of the high-performance concrete detection model based on Mask R-CNN is completed. Simulation experiments are used to analyze the detection and application of high-performance concrete using the theory of machine vision. The data show that the training and validation accuracies of the Mask R-CNN model are 96.75% and 96.52%, respectively, which are 5% and 3.77% higher than that of the migration learning network model (91.75%, 92.75%). In addition, the predictions of the Mask R-CNN model have a relative error of less than 0.05 compared to the actual values. The research presented in this paper is applicable to field strength testing of common concrete structures and can provide theoretical references for the study of construction materials.