Machine Vision Algorithms Research Articles

Deep Neural Network-Based Cigarette Filter Defect Detection System with FPGA Acceleration for Online Recognition

In the cigarette manufacturing industry, machine vision and artificial intelligence algorithms have been employed to improve production efficiency by detecting product defects. However, achieving both high accuracy and real-time defect detection for cigarettes with complex patterns remains a challenge. To address these issues, this study proposes a model based on RESNET18, combined with a feature enhancement algorithm, to improve detection accuracy. Additionally, a method is designed to deploy the model on a field-programmable gate array (FPGA) with high parallel processing capabilities to achieve high-speed detection. Experimental results demonstrate that the proposed detection model achieves a detection accuracy of 95.88% on a cigarette filter defect dataset with an end-to-end detection speed of only 9.38 ms.

Semiconductor Wafer Defect Detection Based on Machine Learning

Semiconductor wafers are widely used in integrated circuit chips. Because of the complex production process, it is easy to cause various defects. Therefore, the defect detection of semiconductor wafers is an important means to ensure their yield and productivity. This paper mainly expounds on the detection method of wafer defects combined with the machine vision algorithm, including the CNN model, and the classification of the learning-based methods. Current problems and future prospects for total crystalline circle defect detection are discussed in the end.

Method of intelligent agricultural pest image recognition based on machine vision algorithm

In recent years, people have paid increasing attention to smart agriculture. Its core technology is to combine information technology such as computer vision with agricultural production. In agricultural production, the occurrence of pests and diseases can seriously affect agricultural production. Therefore, in agricultural production, the diagnosis and identification of agricultural pests and diseases is the key to improve crop yield. Traditional machine vision methods often rely on feature extraction and classifier design, lack of strong learning and generalization ability, and can not effectively control crop diseases and pests in a timely manner. In order to solve this problem, this paper designs an agricultural pest recognition system based on machine vision algorithm of fuzzy recognition theory, and applies it to agricultural pest recognition. Compared with the traditional machine vision method, this method has obvious advantages, and can effectively improve the monitoring and early warning ability of crop diseases and pests. The experimental results show that the highest recognition rate of the machine vision optimization algorithm based on fuzzy recognition theory is 98.06%, and the lowest recognition error rate is 5.83%. It can be seen that the machine vision recognition algorithm based on fuzzy recognition theory has good recognition effect and can be used in agricultural pest recognition system to help farmers carry out agricultural production and reduce economic losses caused by diseases and pests.

Optimization and performance analysis of a novel automatic planting-irrigating integrated robot

Soil desertification is a severe environmental problem that seriously threatens the ecological environment and sustainable development. Traditional tree-planting methods face challenges such as insufficient labor force and low efficiency. Based on these problems, this study designed a novel automatic planting-irrigation integrated robot with automatic planting, irrigating, and monitoring functions. The robot obtains land information through laser triangle ranging technology, the camera, and other sensors. Machine vision and deep learning algorithms are applied for image processing and identification to identify the areas suitable for tree planting accurately. The drip irrigation module adopts the DHT11 digital temperature and humidity sensor control switch, which can realize efficient tree planting and watering operation, and ensure the growth of saplings and has a broad application prospect. Our design is meaningful in restraining soil desertification, maintaining the ecological environment, and realizing a sustainable economy.

Enhanced Defect Management in Strawberry Processing Using Machine Vision: A Cost-Effective Edge Device Solution for Real-Time Detection and Quality Improvement

Managing defects in agricultural fruit processing is crucial for maintaining quality and sustainability in the fruit market. This study explores the use of edge devices, web applications, and machine vision algorithms to improve defect reporting and classification in the strawberry processing sector. A software solution was developed to utilize edge devices for detecting and managing strawberry defects by integrating web applications and machine vision algorithms. The study shows that integrating built-in cameras and machine vision algorithms leads to improved fruit quality and processing efficiency, allowing for better identification and response to defects. Tested in small organic and conventional strawberry processing enterprises, this solution digitizes defect-reporting systems, enhances defect management practices, and offers a user-friendly, cost-effective technology suitable for wider industry adoption. Ultimately, implementing this software enhances the organization and efficiency of fruit production, resulting in better quality control practices and a more sustainable fruit processing industry.

Using machine vision algorithms for characterizing gas-liquid slug flows in vertical pipes

Slug flow, characterized by the distinctive interfacial structures of Taylor bubbles surrounded by liquid films and bridged by aerated liquid slugs, is a dynamically complex two-phase flow pattern exist in many oil and gas, and energy systems. Accurate and precise quantification of such complex flow behaviour is essential for optimal design, safe operation, and reliable modelling of these systems. Existing image-based measurement techniques mostly rely on offline image processing algorithms and are often limited to a narrow set of flow characteristics primarily focusing on Taylor bubbles. Such constraints not only impede real-time flow monitoring and regulation but also leave liquid slug characteristics unmeasured, resulting in an inability to accurately determine the flow characteristics and extract instantaneous void fraction signals. Present study examined the performance of adaptive thresholding (AT) and background subtraction (BS) algorithms in capturing slug flow characteristics. It was found that while the former excels in Taylor bubbles detection and the latter in small bubbles identification, neither individually addresses the accurate measurement of both flow structures' characteristics. This observation, along with the mentioned restrictions of existing algorithms are the main reason for developing the present combined machine vision-based algorithm. While unlocking the ability to extract instantaneous void fraction signals, this new approach facilitates online measurement of a wide range of key flow characteristics, including Taylor bubble length, velocity, void fraction, and surrounding liquid film thickness; liquid slug length and void fraction; and slug unit length, void fraction, and frequency. Parallel to the high-speed imaging, time-series void fraction data was collected using two capacitance sensors installed alongside the imaging area on the pipe, providing benchmark data essential for the validation of the new algorithm's accuracy. The comparisons demonstrated a high degree of accuracy and precision for the combined algorithm. Quantitatively, the new algorithm measured key unit cell characteristics with RMS errors ranging from 2 to 10 %, while the BS and AT algorithms exhibited wider RMS error ranges of 8–46 % and 2–53 %, respectively. This underscores the new algorithm's potential as a transformative tool for slug flow analysis.

Detection of Growth Stages of Chilli Plants in a Hydroponic Grower Using Machine Vision and YOLOv8 Deep Learning Algorithms

Vertical indoor farming (VIF) with hydroponics offers a promising perspective for sustainable food production. Intelligent control of VIF system components plays a key role in reducing operating costs and increasing crop yields. Modern machine vision (MV) systems use deep learning (DL) in combination with camera systems for various tasks in agriculture, such as disease and nutrient deficiency detection, and flower and fruit identification and classification for pollination and harvesting. This study presents the applicability of MV technology with DL modelling to detect the growth stages of chilli plants using YOLOv8 networks. The influence of different bird’s-eye view and side view datasets and different YOLOv8 architectures was analysed. To generate the image data for training and testing the YOLO models, chilli plants were grown in a hydroponic environment and imaged throughout their life cycle using four camera systems. The growth stages were divided into growing, flowering, and fruiting classes. All the trained YOLOv8 models showed reliable identification of growth stages with high accuracy. The results indicate that models trained with data from both views show better generalisation. YOLO’s middle architecture achieved the best performance.

The Good, The Bad and The Ugly: An Open Image Dataset for Automated Sorting of Good, Bad, and Imperfect Produce Using AI and Robotics

In the face of the impending challenge of feeding a growing global population, one-third of all food produced ends up as waste. A notable contributor to this problem is the wastage of a third of perfectly edible and nutritious fresh produce because they need to meet the high cosmetic standards expected by consumers. Eliminating this wastage of imperfect produce is, therefore, a crucial and sustainable means to increase the food supply for a growing global population. This can be achieved through automated sorting of good, bad and imperfect produce using automation, robotics and machine vision. A prerequisite for such automated sorting is fast and accurate machine vision algorithms for successful differentiation between good, bad and imperfect produce. Training such algorithms requires large image datasets. While much work has gone into collecting images of good and bad produce, to the best of our knowledge, no such dataset exists for imperfect produce items. In this paper, we attempt to fill this gap by developing the first publicly available dataset of good, bad and imperfect produce items. The dataset has been made publicly available on the Harvard Dataverse for use in training machine vision algorithms for sorting good, bad and imperfect produce. It is our hope that this open dataset will contribute to improving research and practice for sorting and saving imperfect produce in the food supply chain.

Comparative analysis of different machine vision algorithms for tool wear measurement during machining

Comparative analysis of different machine vision algorithms for tool wear measurement during machining

Colour difference detection algorithm for warp‐knitted fabric based on image colour appearance model

AbstractDetecting colour differences in warp‐knitting fabric is essential to ensure its quality. Machine vision algorithms are commonly used for automatic colour difference detection. However, the current algorithm directly extracts RGB (red, green, blue) colour stimulus values from the image without considering the image's colour attributes, leading to a significant error in colour difference detection. To solve this issue, a colour appearance model is introduced into the field of colour difference detection for warp‐knitted fabric. Initially, a colour appearance model based on CAM16 is established, and the critical parameters are calculated to obtain the colour appearance attribute of the fabric. The colour difference calculation formula is then constructed in the uniform colour space of CAM16‐UCS. A template selection algorithm based on principal component analysis was designed to select warp‐knitted cloth images with standard colours. Subsequently, a cloth colour difference detection algorithm was developed using the cloth colour profile model. The performance of the colour difference formula based on the colour profile model was evaluated using the PF/3 method. To compare the CIELab, CMC(2:1), and CIEDE2000 colour difference formulas, standardised residual sum of squares was used. The results indicated that the colour difference formula based on the colour‐appearance model is about 5.32% different from the visual colour difference perceived by the human eye. However, it can perform as well as the CMC(2:1) colour difference formula, which is widely used in the textile industry.

Quantitative determination of SO2 flux from industrial chimney through machine vision with plume model verification

Accurate emission rate data are essential for assessing the environmental impact of industrial emissions and calibrating the effectiveness of desulphurization equipment. However, accurate calculation of SO2 emissions is difficult to realize due to the environmental conditions, the time-varying and spatial non-uniformity of pollution source emissions. In this paper, a Dual TV Plus machine vision algorithm is proposed, calculated and validated with the conventional Farneback optical flow algorithm in terms of industrial stack emission rates. In addition, the effect of turbulence on the inversion of SO2 emission under complex environmental conditions is evaluated by combining simulations and outdoor experiments. The results show that the Dual TV Plus algorithm performs better than the Farnebäck algorithm and the Dual TV-L1 algorithm, with 50% and 30% improvement in accuracy and 5% and 8% improvement in immunity to interference, respectively. This work is important for promoting the engineering applications of remote sensing instruments for environmental monitoring.

Autopilot drone in construction: A proof of concept for handling lightweight instruments and materials

Applications of Unmanned Aerial Vehicles (UAVs) in construction have been increased in recent years mainly for purposes of automatically collecting field data. In the market, there is a growing opportunity for using UAVs in material handling. On the other side, vertical transportation in construction sites is a crucial challenge that leads to a drop in productivity efficiency. This paper presents a proof of concept for automating the lifting procedure of lightweight instruments and materials in construction sites. To do so, a platform is developed that could automatically manage material/instrument handling requests via an autopilot drone. The aircraft position is tracked with an onboard GPS module which is not accurate in urban areas. So inaccurate GPS signals could deviate the drone from the predefined path. To tackle this problem, a machine vision algorithm is utilized to calibrate the real-time position of the aircraft. A self-supervised learning algorithm is integrated into the system to handling collision avoidance during the flight. Finally, the developed methods are tested with an UAV in a real world environment to demonstrate the potentials of hiring autopilot drones in construction.

VNVC: A Versatile Neural Video Coding Framework for Efficient Human-Machine Vision.

Almost all digital videos are coded into compact representations before being transmitted. Such compact representations need to be decoded back to pixels before being displayed to humans and - as usual - before being enhanced/analyzed by machine vision algorithms. Intuitively, it is more efficient to enhance/analyze the coded representations directly without decoding them into pixels. Therefore, we propose a versatile neural video coding (VNVC) framework, which targets learning compact representations to support both reconstruction and direct enhancement/analysis, thereby being versatile for both human and machine vision. Our VNVC framework has a feature-based compression loop. In the loop, one frame is encoded into compact representations and decoded to an intermediate feature that is obtained before performing reconstruction. The intermediate feature can be used as reference in motion compensation and motion estimation through feature-based temporal context mining and cross-domain motion encoder-decoder to compress the following frames. The intermediate feature is directly fed into video reconstruction, video enhancement, and video analysis networks to evaluate its effectiveness. The evaluation shows that our framework with the intermediate feature achieves high compression efficiency for video reconstruction and satisfactory task performances with lower complexities.

Digital Model of Automatic Plate Turning for Plate Mills Based on Machine Vision and Reinforcement Learning Algorithm

Plate turning is an essential step in the plate rolling process. The traditional control mode relies on the manual observation of billets and mainly manual operation. Manual plate turning becomes an external disturbance of the automatic control system of plate mills, which reduces the reproducibility and accuracy of the rolling process. The automatic plate turning function is urgently needed to improve the control level of the rolling line. In this paper, based on the improved image processing algorithm, the position and angle information of the billet conversion process are detected in real time, and the real-time processing of detection data in a complex production environment is realized. Based on the change in the billet rotation angle in the actual plate turning process, a mathematical model is constructed to simulate the plate turning process. On this basis, the digital model and optimization algorithm for automatic plate turning based on reinforcement learning are established, and the automatic optimization of plate turning speed and accuracy is completed. The field application of data-driven plate turning systems replaces manual plate turning control. The plate turning angle detection error of the system is ≤2°. The average plate turning time of each billet is greatly shortened compared with manual plate turning mode, and the fastest time can be shortened by more than 1 s, which greatly improves the production efficiency and is of great significance for improving the automatic control level and digital upgrade of plate mills.

CONVOLUTIONAL NEURAL NETWORKS USAGE FOR RASTER IMAGE PROCESSING IN NON-DESTRUCTIVE TESTING

An overview of Russian and foreign literature sources on the use of computer vision technologies and convolutional artificial neural networks in technological operations of visual control is presented. Methods and algorithms review of computer vision and artificial neural network models used by the authors of scientific articles, methods of preprocessing and expanding the initial data set for models training under consideration and verifying the correctness of machine vision algorithms is carried out. Possible approaches for automating the assessment of visual quality control in various fields of industry are highlighted, research results, results of machine vision algorithms, neural network models, their accuracy and efficiency indicators are presented, as well as areas and possible scenarios for the application of the presented algorithmic solutions. Some results of machine vision algorithms, neural network models, their accuracy and efficiency indicators are presented.

Pixel-Level Crack Identification for Bridge Concrete Structures Using Unmanned Aerial Vehicle Photography and Deep Learning

Traditional manual inspection technology has the problems of high risk, low efficiency, and being time-consuming in bridge safety management. The unmanned aerial vehicle (UAV)-based detection technology is widely used in bridge structure safety monitoring. However, the existing deep learning-based concrete crack identification method has great limitations in dealing with complex background and tiny cracks in bridge structures. To address these problems, this study designs a crack pixel-level high-performance segmentation model for bridge concrete cracks that is suitable for UAV detection scenarios using machine vision (MV) and deep learning (DL) algorithms. First, considering the high requirements for the computing performance of the MV-based model for UAV-based detection, the ResNet-18-based lightweight convolutional neural network is used to represent the traditional large-scale backbone network of the pyramid scene parsing network (PSPNet) to develop a high-performance crack automatic identification model. Then, considering that bridge concrete cracks have the characteristics of subtle shapes and complex backgrounds, the spatial position self-attention module is inserted into the PSPNet to improve its detection accuracy. A concrete bridge is used for the case study, and a dataset of cracks in bridge concrete structures collected by UAVs is constructed and used for model training. The experimental results show that the loss function of the developed method in the training process results in a smooth decline, and the developed algorithm achieves the evaluation indicators of 0.9008 precision, 0.8750 recall, 0.8820 accuracy, and 0.9012 IOU on the bridge concrete crack dataset, which are significantly higher than other state-of-the-art baseline methods. In addition, four common UAV bridge detection scenarios, including low light, complex crack forms, high background roughness, and complex background scenes, are used to further test the crack detection ability of the developed crack identification model. The experimental results show that the proposed crack identification method can effectively overcome interference and real-size pixel-level segmentation of crack morphology. In addition, it also achieved a detection efficiency of 35.04 FPS, which shows that the real-time detection ability of the method has good applicability in the UAV detection scene.

Displacement sensing based on microscopic vision with high resolution and large measuring range

AbstractMicroimage strain sensing (MISS) is a novel piston‐type sensor based on microscopic vision. In this study, optical disc slice is used as information carriers to improve MISS. There are multiple pits on the surface of an optical disc. By using machine vision algorithms, the pits can be converted into digital information, making them scales for recording displacements. By this means, we proposed a sensing method that combines high resolution, wide range, and strong robustness. The study measured displacement under different conditions. To address inevitable factors such as pixel drift, and manufacturing errors, corresponding compensation methods were provided. The results show that the measurements closely match those of the linear variable differential transformer, with a resolution of up to 20 nm and a range approaching the sensor size. Despite the sensor's dependence on machine vision, it demonstrates strong resistance to environmental factors such as brightness and angle. Combining compensation methods for pixel drift, and manufacturing errors, this sensor can be well‐applied in various working conditions.

Analyzing Players' Tactical Positions and Movement Trajectories in Soccer Matches Using Machine Vision Algorithms

Tactical positions and movement trajectories in soccer matches using machine vision algorithms involves leveraging computer vision techniques to track players and the ball throughout the game. By processing video feeds from multiple cameras, machine vision algorithms can identify players, recognize their positions on the field, and map out their movement trajectories over time. This analysis provides valuable insights into team formations, player positioning, and tactical strategies employed during matches. Coaches and analysts can use this information to assess team performance, identify patterns in gameplay, and make data-driven decisions to optimize tactics and training regimes. Machine vision algorithms enhance the understanding of soccer dynamics, facilitating more strategic and effective approaches to coaching and gameplay.. This paper presents a comprehensive analysis of player movement, tactical positions, and detection in soccer matches using machine learning algorithms. Leveraging a simulation environment built upon a Recurrent Neural Network (RNN) model trained on historical match data, we investigate the efficacy of the model in accurately predicting future player positions based on contextual features such as player trajectories and ball movement. Furthermore, we explore the model's ability to estimate tactical movements during specific time intervals, providing valuable insights for coaches and analysts in understanding team strategies and adapting to match dynamics. Additionally, we evaluate player detection systems to assess their capabilities and limitations, laying the groundwork for future improvements in player tracking technologies. Our findings underscore the potential of machine learning in soccer analytics, offering actionable insights for enhancing performance analysis, strategic decision-making, and overall understanding of the game. Our results show a high prediction accuracy, with an average Mean Squared Error (MSE) of 0.05. Furthermore, we explore the model's ability to estimate tactical movements during specific time intervals, achieving an overall prediction accuracy of 85%. Additionally, we evaluate player detection systems to assess their capabilities and limitations, achieving a detection accuracy of 90%.

Optimization of Layout Indoor Space and Design Based on Machine Vision Algorithm

In the realm of indoor space design, achieving an optimal layout is crucial for enhancing functionality, efficiency, and user experience. Traditional approaches often rely on manual assessments or simplistic algorithms, lacking the precision and adaptability required for modern architectural challenges. This paper presents a novel framework for optimizing indoor space layout and design leveraging advanced machine vision algorithms. By harnessing the power of computer vision techniques, our methodology offers automated analysis and optimization of indoor environments, considering factors such as spatial utilization, traffic flow, and aesthetic appeal. Through a comprehensive exploration of machine learning and optimization techniques, we demonstrate significant improvements in space efficiency and usability, paving the way for intelligent design solutions in various architectural contexts. This research contributes to the intersection of computer vision and architectural design, offering a paradigm shift towards data-driven and adaptive approaches for optimizing indoor spaces.

Lane detection networks based on deep neural networks and temporal information

In the past few years, the lane detection technique has become a key factor for autonomous driving systems and self-driving cars on the road. Among the various vehicle subsystems, the lane detection module is one of the essential parts of the Advanced Driver Assistance System (ADAS). Conventional lane detection approaches use machine vision algorithms to find straight lines in road scene images. However, it is challenging to identify straight or curved lane markings in complex environments. To deal with this problem, this paper presents a lane detection technique based on deep learning. It is combined with a 3D convolutional network, so the temporal information is added to the network architecture. Using the front camera images, the system can immediately detect the lane marking information ahead. Moreover, we propose an approach for improvement by adding the time axis to the network architecture. In addition to using 3D-ResNet50, the temporal convolution and spatial convolution are separated for processing. The accuracy is 91.34% improved after adding time and space split convolution with LeakyReLU. The experiments carried out using real scene images have demonstrated the feasibility of the proposed technique for applications to various complex scenes.