Image Transformation Algorithm Research Articles

Schlieren measurements of shock train flow fields in a supersonic cylindrical isolator at Mach 2

In a supersonic cylindrical isolator at Mach 2, the structures and frequency characteristics of shock train flow fields were experimentally studied by the schlieren measurement method. According to the design principle of parallel light through schlieren windows in a cylindrical duct, a high-precision conformal optical window pair was designed and integratively processed before. Based on a self-built pipeline structure with conformal windows in a direct-connect wind tunnel under adjustable back-pressure conditions, the shock surfaces in a cylindrical isolator at Mach 2 were first captured by the schlieren method. Then, the schlieren photographs were corrected by a nonlinear image transformation algorithm for the restoration of real shock train structures, and the experimental results were compared with numerical simulation results quantitatively. Finally, the shock train positions were calculated by an image recognition algorithm to analyze the self-excited oscillation frequency characteristics of shock train structures. The methods and experiments in this study enriched optical observation methods of supersonic flows through non-rectangular cross-section isolators in scramjet.Graphical abstract

Analysis of Artificial Intelligence Technology and Its Application in Improving the Effectiveness of Physical Education Teaching

To promote the construction of public physical education online courses in colleges and universities and the evaluation of the effectiveness of course teaching, this article combines 3D reconstruction techniques in computer vision to construct a set of human body shape reconstruction models and apply them to physical training exercises and teaching effectiveness assessment tasks. Specifically, first, the joint point location information of the human body in the input image is extracted using the human skeleton analysis algorithm, and modeling the foreground and background pose information of the target region using the Pix2Pix image transformation algorithm; second, multi-scale features such as nodal location features, foreground and background features, high-resolution detail features, and low-resolution global features are fused and the extracted multi-scale features are also decoded with the help of pixel-aligned implicit functions to generate a 3D model of the human body representing the human form.

Salt pond detection on Sentinel 2 Satellite Imagery using Object-Based Image Analysis (OBIA) Approach

Object-based image analysis (OBIA) is an image classification that is oriented to object patterns that use image objects as the basis for processing, calculates characteristics per object, and extracts land cover information from remotely sensed images. This study aims to detect salt ponds using Sentinel 2 satellite data with an object-based classification model. The center of salt production, which is also an experimental area for the development of industrial salt from the ministry of maritime affairs and fisheries on the north coast of the island of Java was selected as the study area. The unit of analysis for this classification is the segmented object of sentinel image. The classification scheme built to detect salt ponds using OBIA consists of level 1, level 2, and level 3. Level 1 is to separate land and water using a Near Infrared canal. Level 2 is to separate land use from object segmentation results in land class at level 1 using NDVI transformation, and level 3 is to separate salt and non-salt ponds from the segmentation results of land use at level 2 using sentinel image transformation algorithm for the distribution of chlorophyll-a. The result shows chlorophyll-a estimation image transformation from sentinel useful to separate salt and non-salt ponds. Many researchers have been reported that chlorophyll-a does not live in the salinity range of salt ponds greater than 50 ppt, meanwhile, in non-salt ponds, chlorophyll-a is used as natural feed for cultivated animals. Furthermore, the research shows a classification scheme of salt ponds and non-salt ponds can be derived from sentinel 2 imagery with OBIA approach

Diagnosis of rotor demagnetization and eccentricity faults for IPMSM based on deep CNN and image recognition

Permanent magnet synchronous motors (PMSMs) have many advantages attributed to the structure of the permanent magnet, which connects the influences of each type of fault. However, diagnostic methods are not capable of monitoring multitype faults with one index or method, due to closed-loop control or non-stationarity. There are a few diagnostic methods based on convolutional neural networks (CNNs) and vibration analysis; however, they are limited by the installed sensor locations and are affected by various factors of the driven system. This paper proposes a method to diagnose eccentricity and demagnetization in an interior PMSM (IPMSM) with image recognition based on a deep CNN and motor stator current analysis. To extract fault features from the IPMSM’s stator current with a deep CNN, a gray image transformation algorithm that uses the autocorrelation matrix is proposed to enhance the feature representations of stator currents, in which currents data are processed recursively and timely. The testing accuracy of 98.74% of both designed model pyramidal Resnet-9 and Resnet-15 indicates that the proposed method is capable of monitoring multitype faults and is immune to speeds and loads.

ПАРАМЕТРИЗАЦИЯ ИЗОБРАЖЕНИЯ ШЕРОХОВАТОЙ ПОВЕРХНОСТИ ДЕТАЛЕЙ МАШИНОСТРОЕНИЯ

Using the transformation of a raster image of a rough surface, combinatorial parameters are introduced, which are determined on the basis of the histogram of the controlled surface fragment. The proposed parameters characterize the area of each level of microroughness. The dynamics of changes in a rough surface as a result of technological impact on it is estimated using the efficiency of filling and the economy of coating. For parameterization of the controlled fragment of the surface, the features of the raster image transformation algorithm were also used. For the image of the GTE blade feather at different stages of its finishing, the introduced parameters for the image of a rough surface were determined. The limits of applicability of the proposed parameters for controlling the dynamics of changes in a rough surface are shown. The ways of developing the proposed parameters for rough surfaces operating under various conditions are indicated.

Measurement of Axial Orientation of Film Cooling Holes Based on a Laser Sensor

The axial orientation of a film cooling hole of an aero-engine turbine blade is one of the main factors affecting the cooling effect. However, due to the characteristics of free-form surfaces, small diameters of cooling holes and the machining error of EDM, measurement of the axial orientation of a film cooling hole is quite difficult on a manufacturing site. Many turbine blade manufacturing companies still use visual judgement assisted by gauge rods manually. The manual inspection is problematic and is not fully capable of giving accurate and reliable results. Thus, this paper proposes a method for measuring axial orientations based on an improved Gaussian image algorithm. First, a point cloud is obtained through a laser scanning sensor. Then, to address the problem of low accuracy of normal vector in traditional Gaussian image algorithms, small number of measured point clouds, and poor planar fitting accuracy, a Gaussian image transformation algorithm based on normal vector characteristics is proposed, which takes cylindrical generatrices into consideration. Further, an augmented normal vector algorithm and a planar fitting algorithm are also included. Finally, based on Geomagic and actual detection experiments, the algorithm is verified for its feasibility, accuracy, and reliability. Experimental results show that the extraction accuracy of the algorithm in the axial orientation can reach 0.6304 °.

Alteration Detection of Multispectral/Hyperspectral Images Using Dual-Path Partial Recurrent Networks

Numerous alteration detection methods are designed based on image transformation algorithms and divergence of bi-temporal images. In the process of feature transformation, pseudo variant information caused by complex external factors will be highlighted. As a result, the error of divergence between the two images will be further enhanced. In this paper, we propose to fuse the variability of Deep Neural Networks’ (DNNs) structure flexibly with various detection algorithms for bi-temporal multispectral/hyperspectral imagery alteration detection. Specifically, the novel Dual-path Partial Recurrent Networks (D-PRNs) was proposed to project more accurate and effective deep features. The Unsupervised Slow Feature Analysis (USFA), Iteratively Reweighted Multivariate Alteration Detection (IRMAD), and Principal Component Analysis (PCA) were then utilized, respectively, with the proposed D-PRNs, to generate two groups of transformed features corresponding to the bi-temporal remote sensing images. We next employed the Chi-square distance to compute the divergence between two groups of transformed features and, thus, obtain the Alteration Intensity Map. Finally, threshold algorithms K-means and Otsu were, respectively, applied to transform the Alteration Intensity Map into Binary Alteration Map. Experiments were conducted on two bi-temporal remote sensing image datasets, and the testing results proved that the proposed alteration detection model using D-PRNs outperformed the state-of-the-art alteration detection model.

Оценка скорости и сложности алгоритмов спектрального преобразования изображений в системах наблюдения за морскими судами

Исследуется процесс развития современных систем видеонаблюдения на морском транспорте, а также затронуты некоторые приложения и особенности передачи изображений и способы повышающие ее эффективность. Системы морского наблюдения могут использоваться для повышения безопасности портов, аэропортов, торговых и военных судов, а также для контроля морского движения в портах и каналах, защиты прибрежных и нефтяных платформ. Камеры являются одним из основных датчиков этих систем. Они дешевы и дополняют другие типы датчиков. В данной работе представлен результаты исследований по использованию алгоритм быстрого преобразования Фурье с децимацией во времени при размерности-22´22 при обработке морских сюжетов, полученных из различных камер наблюдения за крымским мостом. Предложенный алгоритм получен путем применения двухэтапного подхода к декомпозиции и внедрению эффективной методики группировки поворотных множителей Фурье-преобразования в комплексной форме. Анализируется арифметическая сложность предлагаемого алгоритма и вычисляется количество действительных умножений и сложений для различных размеров преобразования и изображений морских судов. Кроме того, выполнена оценка скорости передачи и сложности обработки морских изображений для различных форматов и разрешения. The process of development of modern video surveillance systems in maritime transport is investigated, as well as some applications and features of image transmission and methods that increase its efficiency are touched upon. Maritime surveillance systems can be used to improve the security of ports, airports, commercial and military vessels, as well as to control sea traffic in ports and channels, protect coastal and oil platforms. Cameras are one of the main sensors of these systems. They are cheap and complement other types of sensors. This paper presents the results of research on the use of the fast Fourier transform algorithm with decimation in time at a dimension of-22´22 when processing marine scenes obtained from various surveillance cameras for the Crimean bridge. The proposed algorithm is obtained by applying a two-stage approach to decomposition and implementing an effective method for grouping the rotary multipliers of the Fourier transform in a complex form. The arithmetic complexity of the proposed algorithm is analyzed and the number of real multiplications and additions for various sizes of transformation and images of sea vessels is calculated. In addition, the estimation of the transmission speed and complexity of processing marine images for various formats and resolutions was performed.

Ensemble Method of Deep Learning, Color Segmentation, and Image Transformation to Track, Localize, and Count Cotton Bolls Using a Moving Camera in Real-Time

HighlightsAn ensemble method using color segmentation, deep learning, and image transformation was developed.Experiments were conducted to compare the method with other state-of-the-art tracking algorithms.The optimized ensemble method to track bolls achieved 94.4% accuracy using weakly trained tiny YOLOv2 models.The method achieved 7.6 frames per second and outperformed five other tracking methods.Abstract. In robotic applications, good perception can be computationally costly and create undesirable latency before a control decision is initiated. Most of the methods available for object detection deep learning are either fast with low accuracy or slow with high accuracy. Fast and accurate methods are necessary to track and localize objects such as cotton bolls that may be visible or occluded by each other or not well illuminated. In this study, an ensemble of a deep learning method and other image processing techniques was used to detect cotton bolls in-field on defoliated plants. In each image, a trained deep learning method, the YOLOv2 model, was used to detect open cotton bolls, and color segmentation was applied to confirm if the bolls detected by the YOLOv2 model were actually white to avoid false positives. Boll tracking was performed by following the spatial movement of good features on the edges of the bolls using the Lucas-Kanade algorithm. An image transformation algorithm was applied to the next image in case the previously detected boll was lost to retrieve the information of the missing boll. Each tracked and localized boll was stored and counted to give the total number of bolls detected. In this study, detection accuracy was sacrificed for image processing speed by using the YOLOv2 model. Detection accuracy was improved by using an ensemble method that combined image color segmentation, optical flow, and image transformation. This method was compared to eight other open-source methods implemented in OpenCV. The ensemble method detected and counted bolls at a speed of 7.6 fps with an accuracy of 94.4% using the Jetson TX2 embedded system to process 1K resolution images, outperforming the other OpenCV methods in various measurements. Keywords: Boll counting, Cotton, Cotton harvesting, DarkFlow, Darknet, Deep learning, Machine vision, YOLOv2.

PRZEGLĄD WYKORZYSTANIA AOI W PROCESIE KONTROLI MONTAŻU POWIERZCHNIOWEGO

Technologia montażu powierzchniowego jest obecnie szeroko stosowana w produkcji zespołów obwodów drukowanych w przemyśle elektronicznym. Zyskała ona bardzo wielu zwolenników. Miniaturyzacja komponentów elektronicznych wymusiła wprowadzenie maszyn wizualnej kontroli poprawności montażu, bardziej dokładnych i szybszych niż ludzkie oko, lupa czy mikroskop. Automatyczna Inspekcja Optyczna (AOI) to proces kontroli wykrywania wad i błędów w początkowym procesie produkcji obwodów drukowanych. Staje się nieodzownym elementem montażu kontraktowego, wpływając na zwiększenie jakości oferowanych usług i efektywności produkcji. Wykorzystywane są w niej nowe konstrukcje głowic pomiarowych, miniaturyzacja sprzętu, oprogramowanie przetwarzące otrzymane obrazy płytek, skomplikowane algorytmy przekształcania obrazu.

Transformation of Multispectral Data to Quasi-Hyperspectral Data Using Convolutional Neural Network Regression

Hyperspectral (HS) data are proven to be more resourceful compared to multispectral (MS) data for object detection, classification, and several other applications. However, absence of any space-borne HS sensor since 2017, which can provide open-source data with global coverage, and high cost and limited obtainability of airborne sensors-based images limit the use of HS data. Transformation of readily available MS data into quasi-HS data can be a feasible solution for this issue. In this article, we propose the use of convolutional neural network regression (CNNR), a deep learning-based algorithm, for MS (i.e., Landsat 7/8) to quasi-HS (i.e., quasi-Hyperion) data transformation. The proposed CNNR model is compared with the existing pseudo-HS image transformation algorithm (PHITA), a simple linear model [i.e., stepwise linear regression (SLR)], and a nonlinear modeling approach [i.e., support vector regression (SVR)] by evaluating the quality of the quasi-Hyperion data. Contrary to these existing and simple models, the proposed CNNR model has the added advantage of utilizing deep learning-based spectral-spatial features for MS to quasi-HS data transformation through regression-based nonlinear modeling. Different statistical metrics are calculated to compare each band's reflectance values as well as spectral reflectance curve of each pixel of the quasi-Hyperion data with that of the original Hyperion data. The developed models and generated quasi-Hyperion data are also evaluated with application to crop classification. Analyzing the results of all the experiments, it is evident that CNNR model is more efficient compared to PHITA, SLR, and SVR in creating the quasi-Hyperion data and this transformed data are proven to be resourceful for crop classification application. The proposed CNNR model-based MS to quasi-HS data transformation approach can be used as a viable alternative for different applications in the absence of original HS images.

High capacity partial reversible data hiding by hamming code

A high capacity partial reversible data hiding (PRDH) is introduced in this paper. First of all, an original image is converted to a cover image by the proposed image transformation algorithm. The image transformation algorithm adopts (7,4) Hamming code and minimal pairwise square error to ensure that the generated cover image is an almost distortion-free original image. The secret bits are embedded into the cover image by flipping and modifying the cover bits with respect to the syndrome generated by Hamming code. When the secret bits are extracted from the stego image, it can be transformed back to a cover image by the error-correcting ability provided by Hamming code. And this is the so-called partial reversible property. The visual performance and embedding capacity of the proposed method are theoretical analyzed. According to the experimental and theoretical results, high embedding capacity with acceptable visual performance is achieved by the proposed method. More specifically, the embedding rate is 10.5 times of Jana et al.’s method and Yang et al.’s proposed PRDH, and 3.5 times of Yang et al. ’s modified PRDH.

Implementation of Classic Image Transformation Algorithm to Quantum State, Boundary Extraction and Transformation of Half-Tone Image to Binary

Abstract The aim of the research is computer simulation of a quantum algorithm to solve the problem of transforming a classical image using quantum computing tools and methods, studying recognition algorithms and creating a recognition model using quantum methods. The method of quantum modeling makes it possible to convert a classical image into a quantum state, select boundaries and convert a grayscale image to a binary one, and shows the possibilities of the quantum information theory in interpreting classical problems. The main results of the article are the developed quantum algorithm that allows recognizing objects, as well as the quantum method aimed at representing/processing a color pixel image. The scientific novelty of the article is expressed in the construction of a quantum system, an exponential increase in the speed of solving computational NP-complete problems, which on classical machines can be solved in unacceptable time. The motivation for writing the work was a high growth interest in quantum computing and the benefits that they guarantee. The development of the theoretical foundations of creating software systems and the design of algorithms for new information technologies and specialized computing systems is a dynamic field, as evidenced by the number of existing works in this direction. The developed algorithms for various problems of complexity classes can give a significant gain in efficiency in comparison with existing classical ones and provide a solution to a number of complex mathematical (including cryptographic) problems.

Soot aggregate sizing in an extended premixed flame by high-resolution two-dimensional multi-angle light scattering (2D-MALS)

The spatial distribution of soot aggregate size and morphology within a premixed flat flame (McKenna-type burner and ethyne–air mixture at an equivalence ratio of Φ = 2.7) is characterized by two-dimensional multi-angle light scattering (2D-MALS). A profound investigation of such an extended, radially symmetrical sooting flame with 2D-MALS requires a sophisticated camera calibration to correct for non-linear image scaling and a careful evaluation of the scattering data. Sharp scattering images were acquired in the angular range from 20° to 155° using a rotatable camera system and an automated Scheimpflug adapter. To correct for non-linear variations in horizontal and vertical image magnification occurring at scattering angles differing from perpendicular view, a polynomial-based image transformation algorithm was developed to convert all scattering images into a common coordinate system. Effective radii of gyration and fractal dimensions of soot aggregates were then derived from scattering data by two different approaches. Due to limited amount of angular positions, the classical method based on Guinier and power law analysis shows limitations, as it yields discontinuous results, predominantly in axial direction of the burner. Bayesian analysis was then used for a data fit of the complete structure factor conducting a least square minimization leading to more consistent results. The use of prior knowledge in the Bayesian evaluation allows for improved data fitting and reduced uncertainties in radius of gyration and fractal dimension even for small aggregate sizes.

Design and Implementation of Digital Image Transformation Algorithms

Design and Implementation of Digital Image Transformation Algorithms

Comparison of hyperspectral transformation accuracies of multispectral Landsat TM, ETM+, OLI and EO-1 ALI images for detecting minerals in a geothermal prospect area

Hyperspectral remote sensing generally provides more detailed spectral information and greater accuracy than multispectral remote sensing for identification of surface materials. However, there have been no hyperspectral imagers that cover the entire Earth surface. This lack points to a need for producing pseudo-hyperspectral imagery by hyperspectral transformation from multispectral images. We have recently developed such a method, a Pseudo-Hyperspectral Image Transformation Algorithm (PHITA), which transforms Landsat 7 ETM+ images into pseudo-EO-1 Hyperion images using multiple linear regression models of ETM+ and Hyperion band reflectance data. This study extends the PHITA to transform TM, OLI, and EO-1 ALI sensor images into pseudo-Hyperion images. By choosing a part of the Fish Lake Valley geothermal prospect area in the western United States for study, the pseudo-Hyperion images produced from the TM, ETM+, OLI, and ALI images by PHITA were confirmed to be applicable to mineral mapping. Using a reference map as the truth, three main minerals (muscovite and chlorite mixture, opal, and calcite) were identified with high overall accuracies from the pseudo-images (>95% and >42% for excluding and including unclassified pixels, respectively). The highest accuracy was obtained from the ALI image, followed by ETM+, TM, and OLI images in descending order. The TM, OLI, and ALI images can be alternatives to ETM+ imagery for the hyperspectral transformation that aids the production of pseudo-Hyperion images for areas without high-quality ETM+ images because of scan line corrector failure, and for long-term global monitoring of land surfaces.

Transformation of Landsat imagery into pseudo-hyperspectral imagery by a multiple regression-based model with application to metal deposit-related minerals mapping

Hyperspectral remote sensing is superior to traditional multispectral remote sensing in detailed spectral information but has limited spatial and temporal coverage. Those limitations require an innovative technique that can simulate hyperspectral imagery from multispectral imagery with global coverage, continuous acquisition, and a small number of bands. For this, a combination of Hyperion and Landsat 7 ETM+ images is a representative target. The present study develops a new method, Pseudo-Hyperspectral Image Transformation Algorithm (PHITA), for transforming Landsat 7 ETM+ imagery into pseudo-Hyperion imagery using correlations between Landsat and Hyperion band reflectance data. Each correlation is defined as a multiple linear regression model selected through Bayesian model averaging, in which Hyperion and Landsat bands are dependent and predictor variables, respectively. The resultant pseudo-image has a number of high-quality Hyperion bands of the same scene size as a Landsat image. Through verification of transformation accuracy by statistical analyses and surface mineral mapping, the pseudo-Hyperion image was proven very similar to the original band reflectances, because of large Pearson’s correlation coefficients (generally > 0.94), small RMS error (mostly < 0.016), high structural similarity, and similar appearance of the color composite image. Using a reference mineral map built from an AVIRIS image and field surveys as ground truth, an advantage of the pseudo-image is clarified for the Cuprite hydrothermal alteration area in the western United States. The identification and mapping accuracies of metal deposit-related minerals were high even in areas outside the original Hyperion scene. Featured absorptions were reconstructed in pseudo-reflectance spectra of the typical minerals in the area. The results can enhance the potential of a large Landsat-series dataset over the long term by transformation into pseudo-Hyperion images for global land surfaces.

An efficient brain tumor segmentation based on cellular automata and improved tumor-cut algorithm

Over the last few decades, segmentation applied to numerous applications using medical images have rapidly been increased, especially for the big data of magnetic resonance (MR) images. Brain tumor segmentation on MR images is a challenging task in clinical analysis for surgical and treatment planning. Numerous brain tumor segmentation algorithms have been proposed. However, they have still faced the problems of over and under segmentation according to characteristics of ambiguous tumor boundaries. Improving segmentation method is still a challenging research. This paper presents a framework of two paradigms to improve the brain tumor segmentation; image transformation and segmentation algorithm. To cope with ambiguous tumor boundaries, the proposed novel gray-level co-occurrence matrix based cellular automata (GLCM-CA) is presented. GLCM-CA aims to transform an original MR image to the target featured image. It enhances features of the tumor similar to the background areas prior to segmentation. For segmentation, the efficient Tumor-Cut algorithm is improved. Tumor-Cut is an efficient algorithm in tumor segmentation, but faces the problem of robustness in seed growing leading to under segmentation. To cope with this problem, the novel patch weighted distance is proposed in the proposed Improved Tumor-Cut (ITC). ITC significantly enhances the robustness of seed growing. For performance evaluation, BraTS2013 benchmark dataset is empirically experimented throughout in comparison with the state-of-the-art methods using dice quantitative evaluation metrics. Experiments are carried out on 55 real MR images consisting of training and testing datasets. In this regard, the proposed method based on GLCM-CA feature space and ITC provides the outstanding result superior to the state-of-the-art compared methods.

Learning-Based Image Transformation Algorithm for Legibility Enhancement of Blurred Character

Many people who have visual disturbance, called as “low vision”, such as old sight and cataract feel inconvenient in daily life. For example, climbing stairs, walking the street, reading information on a newspaper or a book, etc.. Especially, they have a difficulty to read characters without using optical aids. In this paper, we propose a novel algorithm to transform a character appearance for compensating the differences of visual perception such as blurring. In the proposed method, characters in an image are transformed into the characters which can be recognized with ease according to pre-learning rules. In the experiments, we apply the proposed method to various characters and fonts. The experimental results show the effectiveness of proposed method.

Patient-specific Modeling of the Heart: Estimation of Ventricular Fiber Orientations

Patient-specific simulations of heart (dys)function aimed at personalizing cardiac therapy are hampered by the absence of in vivo imaging technology for clinically acquiring myocardial fiber orientations. The objective of this project was to develop a methodology to estimate cardiac fiber orientations from in vivo images of patient heart geometries. An accurate representation of ventricular geometry and fiber orientations was reconstructed, respectively, from high-resolution ex vivo structural magnetic resonance (MR) and diffusion tensor (DT) MR images of a normal human heart, referred to as the atlas. Ventricular geometry of a patient heart was extracted, via semiautomatic segmentation, from an in vivo computed tomography (CT) image. Using image transformation algorithms, the atlas ventricular geometry was deformed to match that of the patient. Finally, the deformation field was applied to the atlas fiber orientations to obtain an estimate of patient fiber orientations. The accuracy of the fiber estimates was assessed using six normal and three failing canine hearts. The mean absolute difference between inclination angles of acquired and estimated fiber orientations was 15.4 °. Computational simulations of ventricular activation maps and pseudo-ECGs in sinus rhythm and ventricular tachycardia indicated that there are no significant differences between estimated and acquired fiber orientations at a clinically observable level.The new insights obtained from the project will pave the way for the development of patient-specific models of the heart that can aid physicians in personalized diagnosis and decisions regarding electrophysiological interventions.