Image Acquisition System Research Articles

A method of citrus epidermis defects detection based on an improved YOLOv5

Achieving intelligent detection of citrus epidermal defects after harvesting is of great significance to citrus quality and value assurance. The uneven illumination makes it challenging to detect citrus epidermis abnormalities with great accuracy. In order to improve the detection accuracy of citrus epidermal invisible defects, firstly, a dual-lamp image acquisition system is designed and used to complete the image acquisition of citrus fruit invisible defects. Secondly, the YOLOv5 model was optimized by integrating the attention mechanism CBAM and modifying the loss function as DIoU. Finally, the performance of the improved model was verified by comparison experiments and ablation experiments. According to the experimental results, mAP, Precision and Recall of the improved YOLOv5 model were 95.5%, 94.0% and 95.1%, respectively, which were 5.8%, 3.6% and 7.6% higher than those of YOLOv5x. Meanwhile, the average detection speed increased by 22.1 ms per pic. This indicates that the improved network being applied to citrus epidermal defects detection can achieve better performance. It can provide technical support for the intelligent detection and grading of postharvest citrus.

Crack propagation in viscoelastic finite-sized solids: theory and experiments

The problem of crack propagation in viscoelastic materials is of great interest given the numerous engineering applications of such materials. Due to viscoelasticity, even the study of the basic Mode I opening represents a tricky theoretical challenge. Indeed, existing theories adopt important approximations such as i) simplistic constitutive behaviour, ii) steady-state crack propagation, iii) infinite domain of the system. In this work, we revise the theory of Persson & Brener for systems of infinite domain; specifically, we propose a solution to take into account size effects in a viscoelastic plate. The theory allows to consider the realistic constitutive behaviour of viscoelastic materials and to predict the dependence of the energy release rate with the crack tip speed. Comprehensive experimental investigations are performed to corroborate our theoretical predictions. First, dynamic mechanical analysis (DMA) is performed to characterize the complex viscoelastic modulus of PolyTetraFluoroEthylene (PTFE). Second, tensile tests are carried out on cracked PTFE samples, and pictures are recorded with an image acquisition system. Moreover, a point tracking algorithm is developed to measure the crack length and opening displacement. Moving from small to high crack tip speeds, the fracture process becomes less ductile and an increase in the maximum load is observed. In addition, experimental data show that the inclusion of finite-size effects in the theory is crucial for accurately estimating the energy release rate.

Effects of particle type on the particle fluidization and distribution in a liquid–solid circulating fluidized bed boiler

• Circulating fluidized bed technology is combined with boiler. • Effects of particle type on particle fluidization and distribution are investigated. • Particles accumulation exists at the bottom of the drum for all types of particles. • Particle properties obviously affect the fluidization and distribution of particles. • The non-uniform degree decreases with increased feedwater flow rate. A liquid–solid circulating fluidized bed boiler is designed and built for visualization research by applying the fluidized bed heat transfer and fouling prevention technology to the water side of the boiler. Four types of engineering plastic particles with different physical properties are selected as the solid working media. The effect of particle types on the fluidization and distribution of particles in the boiler is investigated under different feedwater flow rates and amount of added particles by using the charge couple device image measurement and acquisition system. The results show that all kinds of particles can’t be normally fluidized and accumulate in the drum at low amount of added particles and feedwater flow rate. The particles with great density and low sphericity are more likely to accumulate. The average solid holdup in the riser tubes increases with the increase in feedwater flow rate and the amount of added particles. The non-uniform degree of particle distribution in the riser tubes generally decreases with the increase in feedwater flow rate and the amount of added particles. The particles with small density and settling velocity have high average solid holdup in the riser tubes under close sphericity. In generally, the smaller the density and settling velocity, the more uniform the particle distribution in the riser tubes. Three-dimensional diagrams of the non-uniform degree of particle distribution in the riser tubes of the boiler are established.

Experimental Study of the Microscopic Visualization of Gas Clogging during Groundwater Recharge

Clogging is one of the most important factors that restricts the development and popularization of artificial groundwater recharge technology. Gas clogging is an important but often overlooked form of clogging. In this study, a high-speed image acquisition system was used to obtain high-resolution images of the migration of water and gas in the pore. The bypass flow, trapped bubbles in the H-shaped pore channel, blind end, and corner of the pore were directly observed and their clogging mechanisms were analyzed. The influences of the pore structure and gas content on the degree of gas clogging were quantified. The pore–throat size has a certain controlling effect on the movement of the gas and liquid phases. As the diameter of the pore–throat increases, the clogging effect of the gas decreases, and the relative permeability of the water (krw) increases. The pore–throat ratio exhibits a negative correlation with the relative permeability of the liquid phase, and the pore–throat sorting coefficient exhibits a positive correlation with krw. As the gas content increases, the degree of gas clogging increases, and the effect is more significant at low gas-to-liquid ratios (<1:2). These results provide theoretical support for the scientific quantitative evaluation and prediction of the occurrence of gas clogging in groundwater recharge projects.

WHAT IMAGING IS REQUIRED TO PLAN TKA? A COMPARISON BETWEEN EOS AND CT SCAN TO ASSESS CORONAL ALIGNMENT

The emergence of patient specific instrumentation has seen an expansion from simple radiographs to plan total knee arthroplasty (TKA) with modern systems using computed tomography (CT) or magnetic resonance imaging scans. Concerns have emerged regarding accuracy of these non-weight bearing modalities to assess true mechanical axis. The aim of our study was to compare coronal alignment on full length standing AP imaging generated by the EOS acquisition system with the CT coronal scout image.Eligible patients underwent unilateral or bilateral primary TKA for osteoarthritis under the care of investigating surgeon between 2017 and 2022, with both EOS X-Ray Imaging Acquisition System and CT scans performed preoperatively. Coronal mechanical alignment was measured on the supine coronal scout CT scan and the standing HKA EOS.Pre-operative lower limb coronal alignment was assessed on 96 knees prior to TKA on the supine coronal scout CT scan and the standing HKA EOS. There were 56 males (56%), and 44 right knees (44%). The mean age was 68 years (range 53-90). The mean coronal alignment was 4.7 degrees (SD 5.3) on CT scan and 4.6 degrees (SD 6.2) on EOS (p=0.70). There was a strong positive correlation of coronal alignment on CT scan and EOS (pearson0.927, p=0.001). The mean difference between EOS and CT scan was 0.9 degrees (SD 2.4). Less than 3 degrees variation between measures was observed in 87% of knees. On linear regression for every 1° varus increase in CT HKA alignment, the EOS HKA alignment increased by 0.93° in varus orientation. The model explained 86% of the variability.CT demonstrates excellent reliability for assessing coronal lower limb alignment compared to EOS in osteoarthritic knees. This supports the routine use of CT to plan TKA without further weight bearing imaging in routine cases.

Enhance the frame accumulation accuracy of breast multispectral transmission image by the improved Powell registration algorithm

Multispectral transmission imaging provides a new direction for the early screening of breast cancer. Frame accumulation technology can effectively solve the problem of low grayscale and poor multispectral transmission images quality, but the accuracy of the result will be reduced due to the shift between image frames caused by human respiration and jitter. Therefore, an improved Powell registration algorithm based on normalized mutual information is proposed to register sequence images to improve the accuracy of images after frame accumulation. Firstly, the video is acquired and preprocessed through the designed breast transmission image acquisition system. Then, the coarse registration parameters are obtained by the "Rotation Translation scanning" method, which are used as the initial point of Powell algorithm for fine registration to obtain the registered image sequence. Finally, frame accumulation is performed on the registered images. The registration accuracy and the image quality of frame accumulation after registration are evaluated respectively. The results show that the improved method can raise the registration accuracy of the set floating image to more than 98.9%. This method not only improves the registration accuracy of Powell algorithm, but also solves the influence of jitter on frame accumulation technology, and provides a reference for the recognition and segmentation of breast blood vessels or tumors in multispectral transmission images.

X-ray-acquired imaging and detection radiography system using digital radiography with a DSLR digital camera: preliminary results of a pilot study

This study aimed to develop an X-ray digital radiography (DR) imaging system with a low radiation dose and superior resolution compared to those of the existing computed radiography (CR) and DR image acquisition systems using a complementary metal–oxide-semiconductor (CMOS) sensor-type digital single-lens reflex (DSLR) camera that is capable of adjusting sensitivity. The X-ray system consists of an X-ray generator, a DSLR camera, and a power cable. The X-ray irradiated from the X-ray imaging device was detected on a fluorescence plate located on the upper part of the body, and an image was captured by a digital camera installed below the fluorescence plate. After converting the image into an electric signal inside the pixel, the data are output through a cable to the computer, which is a data-processing device installed outside the imaging system. The spatial resolution using the resolution chart was 3.19 (LP/mm). The plot profile obtained using the ImageJ program showed a spatial resolution difference according to the spatial frequency. In addition, by exposing X-rays to the resolution chart, a plot profile of 1.0–4.86 (LP/mm) was drawn for the surface profile and resolution according to spatial frequency. Using an aluminum step wedge, images from 3.0–18 mm and plot profiles were shown. In the phantom, the hand muscle signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were 13.42 was 7.2, and the bone SNR and CNR were 10.86 and 7.2%, respectively. In conclusion, the X-ray image acquisition system using a CMOS sensor-type DSLR camera has the advantage of acquiring images in real-time with low radiation dose, resolution, and contrast, so it applies to X-ray examination in clinical practice, such as the existing CR and DR image acquisition systems. DSLR image acquisition systems can be deployed in developing countries and provided through teleradiology and telemedicine.

Recent Advances in Contrast-Enhanced Photoacoustic Imaging: Overcoming the Physical and Practical Challenges.

For decades now, photoacoustic imaging (PAI) has been investigated to realize its potential as a niche biomedical imaging modality. Despite its highly desirable optical contrast and ultrasonic spatiotemporal resolution, PAI is challenged by such physical limitations as a low signal-to-noise ratio (SNR), diminished image contrast due to strong optical attenuation, and a lower-bound on spatial resolution in deep tissue. In addition, contrast-enhanced PAI has faced practical limitations such as insufficient cell-specific targeting due to low delivery efficiency and difficulties in developing clinically translatable agents. Identifying these limitations is essential to the continuing expansion of the field, and substantial advances in developing contrast-enhancing agents, complemented by high-performance image acquisition systems, have synergistically dealt with the challenges of conventional PAI. This review covers the past four years of research on pushing the physical and practical challenges of PAI in terms of SNR/contrast, spatial resolution, targeted delivery, and clinical application. Promising strategies for dealing with each challenge are reviewed in detail, and future research directions for next generation contrast-enhanced PAI are discussed.

Convolutional Neural Networks for Classifying Electronic Components in Industrial Applications

Electronic component classification often constitutes the uncomplicated task of classifying a single object on a simple background. It is because, in many applications, a technological process employs constant lighting conditions, a fixed camera position, and a designated set of classified components. To date, there has not been an adequate attempt to develop a method for object classification under the above conditions in industrial applications. Therefore, this work focuses on the classification problem of a particular technological process. The process classifies electronic components on an assembly line using a fixed-mounted camera. The research investigated all the essential steps required to build a classification system, such as image acquisition, database creation, and neural network development. The first part of the experiment was devoted to creating an image dataset utilising the proposed image acquisition system. Then, custom and pre-trained networks were developed and tested. The results indicated that the pre-trained network (ResNet50) attained the highest accuracy (99.03%), which was better than the 98.99% achieved in relevant research on classifying elementary components. The proposed solution can be adapted to similar technological processes, where a defined set of components is classified under comparable conditions.

Mechanical properties and failure law of composite rock containing two coplanar fractures

Composite rocks comprise the rock structures that are commonly used in geotechnical engineering. The fracture configuration has a substantial influence on the mechanical behavior, failure mode, and crack propagation of composite rocks. In this study, we considered a composite rock with two prefabricated coplanar fractures. Through laboratory uniaxial compression tests and using a digital image acquisition system, we systematically studied the effects of different fracture lengths and inclination angles on the mechanical properties and failure characteristics of the rocks. We obtained the following results: 1) during the loading deformation of the rock sample, the peak stress and elastic modulus increased with an increase in the fracture inclination angle and decreased with an increase in the fracture length. The deterioration coefficient k (the ratio of the difference between the peak strength of intact and fractured rock sample to that of intact rock sample) decreased with an increase in the fracture inclination angle and increased with an increase in the fracture length. 2) The failure type of the rock samples was primarily controlled by the fracture inclination angle and material of the two rock types, and the fragmentation degree was primarily controlled by the fracture length. With an increase in the fracture inclination angle, the failure mode of rock sample exhibited the following order of changes leading to failure: a double-Y type (trwo wing and one antiwing cracks appeared on each prefabricated fracture) → double-Z type (two wing cracks appeared on each prefabricated fracture) → Z type (one wing crack appeared on each prefabricated fracture). 3) The type of coalescence of the rock bridge was controlled by the fracture inclination angle and structural plane. The crack positions were primarily affected by the fracture length. 4) At a low fracture inclination angle (α ≤ 30°), the propagation of the microcracks showed aggregated band formation. Above moderate fracture inclination angles (α > 30°), the microcrack aggregation band gradually weakened and expanded in the direction of dispersion.

Enhanced three‐dimensional model reconstruction based on local ternary pattern‐guided fusion of multi‐exposure images

AbstractComputer vision applications usually rely on the features extracted from input images with good visibility. Image acquisition systems may produce degraded images with low contrast or distorted colours. For instance, bad weather (haze, fog) can cause images captured outdoor with low visibility. Image processing algorithms generally assume that the input image is the scene radiance. Haze removal, with the recovery of image radiance, ensures reliable features extracted from images and the image processing algorithm can achieve optimal performance. Inspired by the concept of image dehazing, the authors propose an image enhancement method that can be used to improve the visibility of the images. Each original image is first transformed into multiple exposure images by means of gamma‐correction operations and adaptive histogram equalization. The transformed images are analyzed by the computation of the local ternary pattern. The image is then enhanced, with each pixel generated from the set of transformed image pixels weighted by a function of the local pattern feature. The authors evaluate their proposed method on four benchmark image dehazing datasets. The quantitative results show that our method outperforms many deterministic algorithms and deep learning models. Moreover, the authors investigate the impact of image enhancement on a practical image‐based application—the reconstruction of three‐dimensional (3D) model of survey scene. Accurate 3D model reconstruction depends on high‐quality images. Degraded images will result in large errors in the reconstructed 3D model. Experimentations have been carried out on outdoor and indoor surveys. Our analysis finds that when fed into the photogrammetry software, the images enhanced by the authors’ method can reconstruct 3D scene models with sub‐millimetre mean errors, which are much better than those with the original images. As shown in the visual and quantitative results of 3D model reconstruction, the authors’ proposed method also outperforms other image enhancement methods.

Research on Low-Light Enhancement of Eye Image Dataset Based on DLN

Image acquisition system in the process of image acquisition, due to the influence of various uncontrollable factors, especially under adverse conditions such as indoor lighting and individual cases, image acquisition system often has a low contrast ratio, low dynamic range intensity, dark and bright areas of the image details disappear and other defects. Therefore, it has become a problem to obtain a clear image under the condition of finding the function. In this paper, Deep Lightening Network technology is used to analyze eye images under low-light conditions. Different from traditional methods, this method introduces machine learning method to generate enhanced images using learning models. Through experimental verification and comparison, the proposed algorithm can improve the overall brightness and contrast of the image. Based on 485 eye images, compared with two conventional lighten methods, the average PSNR score was 18.2420dB, and the SSIM score was 0.8011. Compared with the second best method, the PSNR and SSIM are improved by 2.02 and 0.01 respectively. Reduce the influence of uneven illumination, improve image quality and sharpness.

Parameter selection and experimental study of the rock particle crushing effect using an image-based method

When maging is used to detect the crushing effect of rock particles, the selected characterization parameters are important factors affecting the results. The image-based detection system is composed of three parts: an image acquisition system, a storage platform, and a digital image processing system.The influence of loading mode and feeding particle size on rock crushing degree and rock morphology characteristics after crushing are analyzed respectively; The crushing ratio and sand-forming ratio of limestone, limestone and granite under shear and extrusion loads are analyzed. The experimental results show that the crushing ratio and sand formation rate play a key role in the crushing of rocks composed of different materials under shear compression loading. The effect analysis of crushing under the feed particle size of 9.5 mm to 16mm shows that there is a great correlation between edges and corners, roundness and overall contour. It provides a basis for the follow-up research on intelligent mine construction and equipment optimization, and is worthy of further popularization and application.

Accuracy of radiostereometric analysis using a motorized Roentgen system in a pilot study for clinical simulation.

Radiostereometric analysis (RSA) is routinely implemented with two paired Roentgen tubes for three-dimensional (3D) implant migration measurements. A conventional set-up of one stationary tube and one mobile could be time-consuming. Utilizing two customized ceiling-mounted tubes is normally associated with investment costs. Thus, a pilot set-up of a motorized system (single Roentgen source) for radiostereometric image acquisition may be a time-saving and space-efficient alternative. RSA using the motorized system is feasible in this study as a non-synchronized image acquisition technique, however, patient motion may occur and influence the assessment of implant migration. The phantom study aimed to assess accuracy of RSA using the motorized Roentgen system in this in vitro study. Accuracy values of translations and rotations were ±0.29mm and ±0.48° for the single Roentgen source RSA set-up and ±0.26mm and ±0.48° for the conventional RSA set-up. This study was also performed to simulate potential patient motion during exposure intervals between paired image acquisition. RSA using the motorized system is able to implement RSA with acceptable accuracy. In general, RSA with synchronized image acquisition is the gold standard to access in vivo implant migration with the highest accuracy. Patient motion exists in non-synchronized image acquisition techniques and results in RSA-related motion artifacts. Then we introduced what RSA-related motion artifacts are. The uniplanar calibration cage applied in the study has a few fiducial and control markers, and some of the markers were occluded in radiographs. Whereas, the number of markers in the calibration cage is correlated with accuracy of 3D implant reconstruction.

Convolutional neural networks for identification of moving combustion chambers entering a brazing process

Despite the effective application of deep learning to image classification tasks, these techniques are yet to be fully implemented in the industrial environment. In particular in brazing processes, where the high temperatures destroy the majority of the identification systems, product identification is required to allow automatic product traceability. This work addresses this issue in an international company facility producing water heaters, overcoming the lack of information regarding which model of the combustion chamber is starting the brazing process. Relying on deep learning algorithms, a solution is presented to develop and implement a vision system capable of classifying, through convolutional neural networks (CNN), the 16 models of combustion chambers. The system deployed at the company's production line, although trained with a relatively small collection of images, achieved an F1-score weighted average of 96% and 100% accuracy in classifying some of the models of combustion chambers, already implemented on the shop floor. From the building of the vision system for image acquisition until the training and deployment of the algorithms to classifying the combustion chambers are discussed.

Discharge characteristics of positive and negative polarity streamers with submillimeter gap based on fractal theory

In order to study the characteristics of streamer discharge in sub-millimeter gap, an experimental system and an image acquisition system of positive and negative polarity streamer discharge in the gap of 100–500 μm at atmospheric pressure are built, and the initial breakdown voltages and breakdown images of positive and negative polarity streamer discharge in a gap range of 100–500 μm are collected, and the fractal dimension of the experimental discharge images is calculated by Matlab. The experimental analysis shows that the air breakdown discharge in the gap range of 100–500 μm still conforms to the stream theory, and the polarity effect between the breakdown voltage of positive and negative streamers and the fractal dimension of the discharge image is obvious. The influence of space charge and photoionization on the meandering degree and coverage of the discharge channel in stream theory increase with the increase of the gap, and jump at 300 μm. The fractal simulation of positive and negative streamer discharge with a gap of 500 μm is established by using WZ (Wiesmann-Zeller) model based on fractal theory in Matlab. It is found that the development probability index <i>η</i> is inversely proportional to the fractal dimension, and that the <i>η</i> is smaller in positive streamer than in negative streamer under the same discharge condition.

Honey quality detection based on near-infrared spectroscopy

As a natural agricultural product, honey is favored by consumers, and its variety and adulteration have a huge impact on the quality. Acacia honey, red jujube honey and rape honey were used as experimental objects, and their spectral reflectance curves were obtained through a near-infrared spectral image acquisition system. Spectral features were extracted from the preprocessed spectral reflectance curves, and a honey variety classification model based on near-infrared spectral features was established by machine learning. After statistical analysis, Principal Component Analysis Support Vector Machine after processing data through Successive Projections Algorithm (SPA-SVM) is the optimal classification model for three varieties of acacia honey, red jujube honey and rape honey, and the correct rate of honey variety classification reaches 95.83%. The spectral reflectance curve was used to establish a honey adulteration identification model based on the partial least squares-discriiminate analysis (PLS-DA), and the classification accuracy was 97.92% in the test set.

Retracted: Design of Image Acquisition System Based on Visual Intelligent Sensor and Video Signal Processing

Retracted: Design of Image Acquisition System Based on Visual Intelligent Sensor and Video Signal Processing

Design and implementation of Pedestrian detection based on Zynq APSoc

Pedestrian recognition is an important content of intelligent security surveillance video processing system. In order to meet the application requirements of miniaturization and real-time security system, a moving target recognition system based on APSoc platform is designed. The system uses ov5640 camera as the video collector, zynq7020 as the development platform, and uses the soft and hard co-design to realize each function module. Finally, the pedestrian image acquisition and real-time display system is successfully completed. The test results show that the system can effectively and stably track pedestrians and moving objects within a certain range of distance, and can be displayed in real time, miniaturized and low power consumption, which can be further applied in the field of security.

WSVAS: A YOLOv4 -based phenotyping platform for automatically detecting the salt tolerance of wheat based on seed germination vigour.

Salt stress is one of the major environmental stress factors that affect and limit wheat production worldwide. Therefore, properly evaluating wheat genotypes during the germination stage could be one of the effective ways to improve yield. Currently, phenotypic identification platforms are widely used in the seed breeding process, which can improve the speed of detection compared with traditional methods. We developed the Wheat Seed Vigour Assessment System (WSVAS), which enables rapid and accurate detection of wheat seed germination using the lightweight convolutional neural network YOLOv4. The WSVAS system can automatically acquire, process and analyse image data of wheat varieties to evaluate the response of wheat seeds to salt stress under controlled environments. The WSVAS image acquisition system was set up to continuously acquire images of seeds of four wheat varieties under three types of salt stress. In this paper, we verified the accuracy of WSVAS by comparing manual scoring. The cumulative germination curves of wheat seeds of four genotypes under three salt stresses were also investigated. In this study, we compared three models, VGG16 + Faster R-CNN, ResNet50 + Faster R-CNN and YOLOv4. We found that YOLOv4 was the best model for wheat seed germination target detection, and the results showed that the model achieved an average detection accuracy (mAP) of 97.59%, a recall rate (Recall) of 97.35% and the detection speed was up to 6.82 FPS. This proved that the model could effectively detect the number of germinating seeds in wheat. In addition, the germination rate and germination index of the two indicators were highly correlated with germination vigour, indicating significant differences in salt tolerance amongst wheat varieties. WSVAS can quantify plant stress caused by salt stress and provides a powerful tool for salt-tolerant wheat breeding.