Morphological Processing Research Articles

Research on bearing surface defect detection system based on machine vision

Aiming at the shortcomings of manual detection methods for surface defects of bearing outer rings, a surface defect detection method based on machine vision is proposed. The common defects on the bearing surface are scratches, cracks, and peeling. The defect images are analyzed by the LabVIEW image processing module. Make fast and accurate inspections. In this paper, a CMOS industrial camera is used to obtain the bearing image, and the gray histogram of the image is analyzed to determine whether it is a defective bearing. In order to solve the problem of uneven illumination in image segmentation, it is proposed to first use morphological processing for background. Then, the difference map containing defect information is obtained, and then the Otsu method is used for image segmentation; an edge detection method is designed based on morphology, which can achieve complete edge extraction and good denoising effect. Experiments show that the method has strong practicability and self-adaptability, and can accurately detect bearing surface defects.

Contrastive Study of Brain Activation Regions during Verb Production in Uyghur and Chinese Broca Aphasia Patients after Stroke

<bold>Objective</bold> Blood oxygenation level dependent-functional magnetic resonance image (Bold-fMRI) was used to compare the brain activation of Uyghur and Chinese broca aphasia patients during verb production task, exploring the neural mechanism of verb production, in order to provide basis for the evaluation and treatment of Uyghur and Chinese broca aphasia patients. <bold>Methods</bold> A total of thirty-one patients with broca aphasia who met the inclusion and exclusion criteria were selected, including 15 patients in the Uyghur patient group and 16 patients among the Chinese patient group. Thirty-six healthy volunteers were recruited and matched with the patients in the ethnicity, sex, age and years of education, with 18 in each group. Language function was evaluated by Uyghur and Chinese standard aphasia test. Using Uyghur and Chinese nouns as basic experimental materials, Bold-fMRI data were collected when subjects performed the verb production task. SPM 8.0 software was used for statistical analysis, language and aphasia factors were used for double factor variance analysis to compare the interaction at the whole brain level of the four groups. Simple effect analysis was performed in the brain regions with significant interaction, and the differences in brain activation were compared between the two groups. <bold>Results</bold> In terms of language functions, there were no significant differences in repetition, naming, listening and understanding, reading, writing scores and BDAE grading between the Uyghur and Chinese patient groups ( <italic>P</italic>&gt;0.05). The four groups showed significant interaction in the activation of bilateral lateral occipital lobe, middle frontal gyrus, precentral gyrus, superior temporal gyrus and fronto-orbital gyrus; there were statistically significant differences in activation intensity of bilateral middle frontal gyrus between the Uyghur healthy group and patient group ( <italic>P</italic>&lt;0.05), but there were no statistically significant differences in activation intensity of precentral gyrus and left lateral occipital lobe ( <italic>P</italic>&gt;0.05). There were statistically significant differences in activation intensity of precentral gyrus and left lateral occipital lobe between the Chinese healthy group and patient group ( <italic>P</italic>&lt;0.05), while there were no statistically significant differences in activation intensity of bilateral middle frontal gyrus ( <italic>P</italic>&gt;0.05). There were no statistically significant differences in simple effect comparison between the Uyghur healthy group and Chinese healthy group, or the Uyghur patient group and Chinese patient group ( <italic>P</italic>&gt;0.05). <bold>Conclusion</bold> Language type and aphasia both affect neural mechanism of brain verb production; the left middle frontal gyrus may play an important role in morphological processing of Uyghur verbs; the precentral gyrus may be involved in phonetic coding output of Chinese.

Effect of melting temperatures on the orientation and rheology of polyacrylate/silica (PAcr/SiO2) composite microspheres in twin‐screw extruder

AbstractThe dispersion and orientation arrangement of functional polymers with fillers are critical to their performance. However, there are rare detailed researches on the visualization of the orientation process about functional fillers in composite melts and the mechanistic effects on performance. In this work, it explored the change process of the orientation morphology for polyacrylate/silica (PAcr/SiO2) composite microspheres in polymethyl methacrylate (PMMA) at different melting temperatures. Afterwards, the morphology and distribution of oriented microfibers in composite fibers melted at different temperatures were observed by transmission electron microscope (TEM). The results showed that composite microfibers could be formed in tight arrangement and obtained average orientation of 50.88 in the composite fibers at 250°C. Under this circumstance, the addition of nano SiO2 particles had no effect on the rheological type. Meanwhile, the shear stress was 99.29% higher than that of pure PMMA system under the same conditions. As for the elongation at break, it was found that the elongation at break of the composite fiber prepared at 250°C was 169.43% higher than that of pure PMMA. Therefore, it provided a new method to achieve well dispersion of functional fillers in molten systems. Furthermore, it was beneficial to enhance the elongation at break of functional composites.

The Computer-Aided Design Algorithm of Dyeing and Weaving Graphics from the Perspective of Public Art.

With the acceleration of economic development, people put forward higher requirements for clothing style. In this context, the application of traditional patterns has good artistic effects, can show a unique style, and can also express the artistic beauty of clothing through deformation, color distribution, and other forms, and occupies a relatively large position in the design process of dyeing and weaving art. Aiming at the problem of inaccurate extraction of image information from known visual scenes in the original art-aided design, resulting in unclear output images, this paper proposes a computer-aided design algorithm for dyeing and weaving graphics in the field of public art by color segmentation of the known visual scene images according to the set threshold, morphological processing of the segmented images, reducing noise and fractures affecting the acquired connected areas, and formulating extraction rules to screen candidate areas. Furthermore, dense sampling form is used to extract more scale invariant feature transform (SIFT) target features in the candidate area, match the feature points, integrate the coordinate system of known image information into a unified coordinate system, output the design image, and complete the auxiliary design of dyeing and weaving graphics. The results of simulation experiments show that the computer-aided design algorithm of dyeing and weaving graphics in the public art field is more accurate than the original method in extracting information from known images, which helps to solve the problem of clear output dyeing and weaving images, and improves the quality of dyeing and weaving images.

Efficient fetal ultrasound image segmentation for automatic head circumference measurement using a lightweight deep convolutional neural network.

Fetal head circumference (HC) is an important biometric parameter that can be used to assess fetal development in obstetric clinical practice. Most of the existing methods use deep neural network to accomplish the task of automatic fetal HC measurement from two-dimensional ultrasound images, and some of them achieved relatively high prediction accuracy. However, few of these methods focused on optimizing model efficiency performance. Our purpose is to develop a more efficient approach for this task, which could help doctors measure HC faster and would be more suitable for deployment on devices with scarce computing resources. In this paper, we present a very lightweight deep convolutional neural network to achieve automatic fetal head segmentation from ultrasound images. By using sequential prediction network architecture, the proposed model could perform much faster inference while maintaining a high prediction accuracy. In addition, we used depthwise separable convolution to replace part of the standard convolution in the network and shrunk the input image to further improve model efficiency. After getting fetal head segmentation results, post-processing, including morphological processing and least-squares ellipse fitting, was applied to obtain the fetal HC. All experiments in this work were performed on a public dataset, HC18, with 999 fetal ultrasound images for training and 335 for testing. The dataset is publicly available on https://hc18.grand-challenge.org/ and the code for our method is also publicly available on https://github.com/ApeMocker/CSM-for-fetal-HC-measurement. Our model has only 0.13 million [M] parameters and can achieve an inference speed of 28 [ms] per frame on a CPU and 0.194 [ms] per frame on a GPU, which far exceeds all existing deep learning-based models as far as we know. Experimental results showed that the method achieved a mean absolute difference of 1.97( ± 1.89) [mm] and a Dice similarity coefficient of 97.61( ± 1.72) [%] on HC18 test set, which were comparable to the state of the art. We presented a very lightweight deep learning-based model to realize fast and accurate fetal head segmentation from two-dimensional ultrasound image, which is then used for calculating the fetal HC. The proposed method could help obstetricians measure the fetal HC more efficiently with high accuracy, and has the potential to be applied to the situations where computing resources are relatively scarce.

Machine Learning Approach for Application-Tailored Nanolubricants' Design.

The fascinating tribological phenomenon of carbon nanotubes (CNTs) observed at the nanoscale was confirmed in our numerous macroscale experiments. We designed and employed CNT-containing nanolubricants strictly for polymer lubrication. In this paper, we present the experiment characterising how the CNT structure determines its lubricity on various types of polymers. There is a complex correlation between the microscopic and spectral properties of CNTs and the tribological parameters of the resulting lubricants. This confirms indirectly that the nature of the tribological mechanisms driven by the variety of CNT–polymer interactions might be far more complex than ever described before. We propose plasmonic interactions as an extension for existing models describing the tribological roles of nanomaterials. In the absence of quantitative microscopic calculations of tribological parameters, phenomenological strategies must be employed. One of the most powerful emerging numerical methods is machine learning (ML). Here, we propose to use this technique, in combination with molecular and supramolecular recognition, to understand the morphology and macro-assembly processing strategies for the targeted design of superlubricants.

From decomposition to distributed theories of morphological processing in reading.

The morphological structure of complex words impacts how they are processed during visual word recognition. This impact varies over the course of reading acquisition and for different languages and writing systems. Many theories of morphological processing rely on a decomposition mechanism, in which words are decomposed into explicit representations of their constituent morphemes. In distributed accounts, in contrast, morphological sensitivity arises from the tuning of finer-grained representations to useful statistical regularities in the form-to-meaning mapping, without the need for explicit morpheme representations. In this theoretically guided review, we summarize research into the mechanisms of morphological processing, and discuss findings within the context of decomposition and distributed accounts. Although many findings fit within a decomposition model of morphological processing, we suggest that the full range of results is more naturally explained by a distributed approach, and discuss additional benefits of adopting this perspective.

INFLUENCE OF CURRENT DENSITY ON THE PROCESS OF FORMATION OF OXIDE COATING ON ZIRCONIUM ALLOY E110

Today, binary, triple and quadruple zirconium alloys containing iron, chromium and copper are known for corrosion protection, which have a clear advantage over circal and niobium binary alloys in terms of corrosion resistance. However, insufficient attention is paid to studies aimed at studying the process of formation and morphology of oxide films on the zirconium alloy E110. This area of research has a high scientific and technical interest in the development of oxide coatings on metal surfaces, given the widespread use of zirconium alloys in nuclear energy as a structural material in the construction of nuclear reactors. The paper considers the advantages of the method of microarc oxidation in comparison with other methods of protection of zirconium alloys from different types of corrosion damage. The main task of the work is to study the influence of current density on the formation and morphology of oxide films on zirconium alloy. It was found that the obtained dependences of the formation stress on the time of microarc oxidation of zirconium alloy in 1 mol/l potassium pyrophosphate at different current densities have a characteristic classical appearance, characterized by three sections, such as spark, spark and microarc. The dependence of the rate of change of the voltage dU / dt on the molding voltage is constructed for a more detailed study of the process of formation of oxide films and determination of the moment of the beginning of sparking and microarcs. It was found that when the current density increases to 6 A/m2, the spark voltage increases by 5 V, and when it increases to 10 A/m2 - by 15 V, but the voltage of the microarcs remains unchanged. Mostly solid, well-bonded to the base oxide films of light gray colour on a zirconium alloy are obtained. The morphology of oxide coatings obtained on the surface of the zirconium alloy E110 by the method of microarc oxidation in potassium pyrophosphate at different current densities was studied. It is shown that oxide films without cracks and surfaces have irregularities with a cellular texture.

A Method Based on Multi-Network Feature Fusion and Random Forest for Foreign Objects Detection on Transmission Lines

Foreign objects such as kites, nests and balloons, etc., suspended on transmission lines may shorten the insulation distance and cause short-circuits between phases. A detection method for foreign objects on transmission lines is proposed, which combines multi-network feature fusion and random forest. Firstly, the foreign object image dataset of balloons, kites, nests and plastic was established. Then, the Otus binarization threshold segmentation and morphology processing were applied to extract the target region of the foreign object. The features of the target region were extracted by five types of convolutional neural networks (CNN): GoogLeNet, DenseNet-201, EfficientNet-B0, ResNet-101, AlexNet and then fused by concatenation fusion strategy. Furthermore, the fused features in different schemes were used to train and test random forest, meanwhile, the gradient-weighted class activation mapping (Grad-CAM) was used to visualize the decision region of each network, which can verify the effectiveness of the optimal feature fusion scheme. Simulation results indicate that the detection accuracy of the proposed method can reach 95.88%, whose performance is better than the model of a single network. This study provides references for detection of foreign objects suspended on transmission lines.

A developmental perspective on morphological processing in the flankers task

Recent research with adult participants using the flankers task has shown that the recognition of central target words is facilitated by the presence of morphologically related flanker words. Here we explored the development of such morphological flanker effects in two groups of primary school children (average ages = 8 years 6 months and 10 years 3 months) and a group of adult participants. We examined effects of a transparent morphological relation in two conditions: one where the target was the stem and flankers were derivations (e.g., farmer farm farmer) and the other where the flankers were stems and the target was the derived form (e.g., farm farmer farm). Morphological flanker effects were compared with repetition flanker effects with the same set of stimuli (e.g., farm farm farm; farmer farmer farmer), and effects of related flankers were contrasted with the appropriate unrelated flankers. Results revealed no significant effect of morphological relatedness in the two groups of children and a significant effect in the adult group, but only for suffixed targets and stem flankers. Repetition effects for stem targets were found across all groups, whereas repetition effects for suffixed targets were found only in the older children and adults. These results show that morphological processing, in a context involving multiple words presented simultaneously, takes several years to develop and that morphological complexity (stem vs. derived) is a limiting factor for repetition effects in the flankers task with young children.

Noun Composition in Narrative Passage

The morphology process in forming the words needs a formation component or element of words, an affixation process, a reduplication process, and a composition process. Composition is combining two or more words that create a new meaning. The problem focus of this research was basic composition in narrative passages with basic forms of nouns and adjectives. The method used in this research was the descriptive qualitative method. The researcher used this descriptive qualitative method to describe the basic composition form and affixed composition objectively. The data source used in this research was narrative passage research on the origins of the Jombang Regency. Data were in the form of Indonesian language composition. The research found categories of basic noun composition forms such as (N+N) and (N+Adj). The form of the adjective basic composition category found is (Adj + V) and (Adj + Adj).

Solar Hαexcess during Solar Cycle 24 from full-disk filtergrams of the Chromospheric Telescope

Context.The chromospheric Hαspectral line is a strong line in the spectrum of the Sun and other stars. In the stellar regime, this spectral line is already used as a powerful tracer of stellar activity. For the Sun, other tracers, such as Ca IIK, are typically used to monitor solar activity. Nonetheless, the Sun is observed constantly in Hαwith globally distributed ground-based full-disk imagers.Aims.The aim of this study is to introduce the imaging Hαexcess and deficit as tracers of solar activity and compare them to other established indicators. Furthermore, we investigate whether the active region coverage fraction or the changing Hαexcess in the active regions dominates temporal variability in solar Hαobservations.Methods.We used observations of full-disk Hαfiltergrams of the Chromospheric Telescope and morphological image processing techniques to extract the imaging Hαexcess and deficit, which were derived from the intensities above or below 10% of the median intensity in the filtergrams, respectively. These thresholds allowed us to filter for bright features (plage regions) and dark absorption features (filaments and sunspots). In addition, the thresholds were used to calculate the mean intensityImeanE/Dfor Hαexcess and deficit regions. We describe the evolution of the Hαexcess and deficit during Solar Cycle 24 and compare it to the mean intensity and other well established tracers: the relative sunspot number, the F10.7 cm radio flux, and the Mg IIindex. In particular, we tried to determine how constant the Hαexcess and number density of Hαexcess regions are between solar maximum and minimum. The number of pixels above or below the intensity thresholds were used to calculate the area coverage fraction of Hαexcess and deficit regions on the Sun, which was compared to the imaging Hαexcess and deficit and the respective mean intensities averaged for the length of one Carrington rotation. In addition, we present the Hαexcess and mean intensity variation of selected active regions during their disk passage in comparison to the number of pixels of Hαexcess regions.Results.The Hαexcess and deficit follow the behavior of the solar activity over the course of the cycle. They both peak around solar maximum, whereby the peak of the Hαdeficit is shortly after the solar maximum. Nonetheless, the correlation of the monthly averages of the Hαexcess and deficit is high with a Spearman correlation ofρ = 0.91. The Hαexcess is closely correlated to the chromospheric Mg IIindex with a correlation of 0.95. The highest correlation of the Hαdeficit is found with the F10.7 cm radio flux, with a correlation of 0.89, due to their peaks after the solar activity maximum. Furthermore, the Hαdeficit reflects the cyclic behavior of polar crown filaments and their disappearance shortly before the solar maximum. We investigated the mean intensity distribution for Hαexcess regions for solar minimum and maximum. The shape of the distributions for solar minimum and maximum is very similar, but with different amplitudes. Furthermore, we found that the area coverage fraction of Hαexcess regions and the Hαexcess are strongly correlated with an overall Spearman correlation of 0.92. The correlation between the Hαexcess and the mean intensity of Hαexcess regions is 0.75. The correlation of the area coverage fraction and the mean intensity of Hαexcess regions is in general relatively low (ρ = 0.45) and only for few active regions is this correlation above 0.7. The weak correlation between the area coverage fraction and mean intensity leaves us pessimistic that the degeneracy between these two quantities can be broken for the modeling of unresolved stellar surfaces.

Microstructure reconstruction using fiber tracking technique and pore-scale simulations of heterogeneous gas diffusion layer

Two challenging tasks in pore-scale modeling of a gas diffusion layer (GDL) are realistic microstructure reconstruction and stress-strain simulation to differentiate the heterogeneous materials. This study proposes a novel method for reconstructing a GDL using fiber tracking technique and pore-scale modeling to investigate its stress-strain and anisotropic transport properties. X-ray computed tomography, fiber tracking, and morphological processing techniques were employed to reconstruct a realistic GDL. Pore-scale modeling was performed to compute the stress-strain, gas diffusivity, and electrical-thermal conductivity at different compression ratios. The sensitivity of compression speed and Young's modulus were investigated to balance the accuracy and computing cost of stress-strain simulation. The results showed that Young's modulus of 1 GPa and compression speed of 3 m/s meet the requirements for both accuracy and computational cost. The reconstructed GDL showed good agreements with the experimental data when considering fibers' orientation, length, and curvature. It was found that the stress among fibers was approximately five times higher than binders. The anisotropic ratios of diffusivity and conductivity decreased from 1.35 to 1.25, and 15 to 5, respectively, as the compression ratio increased to 25%. This study can provide accurate predictions and guidelines for GDL design with low stress and high performance.

The Effect of Semantic Transparency in a Flanker Task.

This study tried to replicate and extend the semantic transparency morphological effect using the flanker lexical decision paradigm (Grainger et al., 2020). In the first experiment, stems were used as flankers of target words that could be truly morphological (hunt hunter hunt), pseudomorphological (corn corner corn), or form-related with the flanker (broth brothel broth). In half of the trials, a related flanker was employed, and in the other half, an unrelated word was presented as flanker (e.g., table player table). The results showed a facilitative effect for the related condition as a main effect with no difference between experimental conditions. These results were interpreted in terms of an orthographic facilitation taking place when whole stems are presented as flankers. In the second experiment, short derivational suffixes were used as flankers of the same targets employed in the first experiment. The results showed an inhibitory effect of the same magnitude for the transparent and pseudomorphological conditions with no effect for the form condition. This finding suggests an inhibitory effect by which morphemes activate several lexical candidates that compete for recognition. Overall, the results are interpreted in terms of the cognitive requirements of the experimental task, the items selected, and the current models of morphological processing.

Simulation and experimental investigation of the surface morphology formation mechanism of a D-shaped fiber processed using a pulsed CO2 laser

Traditional D-shaped optical fibers are manufactured using mechanical polishing or chemical corrosion techniques, which inevitably introduce sub-surface defects or impurities. In this study, a new D-shaped optical fiber manufacturing method using CO2 laser ablation technology is proposed. Laser ablation technology is a complex multi-physical coupling process involving phase changes. To fully understand the surface morphology formation mechanism during CO2 pulsed laser processing, a three-dimensional numerical model and experimental analyses were used to elucidate the relationship between the fiber surface temperature distribution, surface morphology evolution, processing quality, processing efficiency, and laser parameters. The simulation results showed that the residual temperature between laser pulses was lower than the structural transition temperature of the material, and the simulated morphology was consistent with the experimental results. Moreover, the overlapping spot energy resulted in the processed surface exhibiting a periodic ripple structure. The ablation depth was proportional to the pulse width and inversely proportional to the scanning speed. In addition, different laser pulse widths had almost no effect on the roughness and flatness of the processed surface, but the laser scanning speed had a significant impact on both indexes. These results can guide the optimization of CO2 laser processing methods.

The effect of melt drawing ratio on the crystal structural change and performance of polyketone extrusion cast film

In this paper, the polyketone (POK) extrusion cast film is manufactured by melt stretching method, and the evolution process of the crystal morphology and mechanical properties with the increase of melt drawing ratio (MDR) are followed. The results show that the melt stretching process produces many micro shish-kebab crystals in the POK. The length of the shish crystal and the thickness of the kebab crystal hardly increase with MDR, but the lateral length of the kebab crystal shows linear growth when MDR exceeds 40. The crystalline morphology of POK is mainly affected by melt relaxation. The molecular chain has sufficient relaxation during cooling at a low MDR (20–40). At this time, micro shish-kebab crystals are mainly randomly arranged. When MDR exceeds 40, the rapid melt stretching shortens the relaxation time of the tie chain between the neighborhood shish crystal, and the atomic force microscopy image shows a typical shish-kebab structure. This experimental result indicates that the formation of the oriented lamella structure may be related to the relaxation of the molecular chains between the micro-shish. When the length of the shish axis and the thickness of the kebab lamellae are similar, it is difficult to distinguish the two.

Detection of physical descaling damage in carp based on hyperspectral images and dimension reduction of principal component analysis combined with pixel values.

The surface of carp is easily damaged during the descaling process, which jeopardizes the quality and safety of carp products. Damage recognition realized by manual detection is an important factor restricting the automation in the pretreatment. For the commonly used methods of mechanical and water-jet descaling, damage area recognition according to the hyperspectral data was proposed. Two discrimination models, including decision tree (DT) and self-organizing feature mapping (SOM), were established to recognize the damaged and normal descaling area with the average spectral value. The damage-discrimination model based on DT was determined to be the optimal one, which possessed the best model performance (accuracy = 96.7%, sensitivity = 96.7%, specificity = 96.7%, F1-score = 96.7%). Considering the efficiency and precision of damage-area recognition and visualization, the principal component analysis (PCA) combined with pixel values statistical analysis was used to reduce the dimension of hyperspectral images at the image level. Through statistical analysis, the value 0 was used as the threshold to distinguish the normal area and the damaged area in the PC image to achieve preliminary segmentation. Then, the spectral values of the initially discriminated damage area were input into the DT discrimination model to realize the final discriminant of damaged area. On this basis, the position information of the damaged area could be used to realize the visualization. The final visualization maps for mechanical and water-jet descaling damage were obtained by image morphology processing. The average recognition accuracy can reach 94.9% and 90.3%, respectively. The results revealed that the hyperspectral imaging technique has great potential to recognize the carp damage area nondestructively and accurately under descaling processing. PRACTICAL APPLICATION: This study demonstrated that hyperspectral imaging technique can realize the carp damage area detection nondestructively and accurately under descaling processing. With the advantages of nondestructive and rapid, hyperspectral imaging system and the method can be widely expanded and applied to the quality detection of other freshwater fish pretreatment.

Cotton Yield Estimation From Aerial Imagery Using Machine Learning Approaches.

Estimation of cotton yield before harvest offers many benefits to breeding programs, researchers and producers. Remote sensing enables efficient and consistent estimation of cotton yields, as opposed to traditional field measurements and surveys. The overall goal of this study was to develop a data processing pipeline to perform fast and accurate pre-harvest yield predictions of cotton breeding fields from aerial imagery using machine learning techniques. By using only a single plot image extracted from an orthomosaic map, a Support Vector Machine (SVM) classifier with four selected features was trained to identify the cotton pixels present in each plot image. The SVM classifier achieved an accuracy of 89%, a precision of 86%, a recall of 75%, and an F1-score of 80% at recognizing cotton pixels. After performing morphological image processing operations and applying a connected components algorithm, the classified cotton pixels were clustered to predict the number of cotton bolls at the plot level. Our model fitted the ground truth counts with an R2 value of 0.93, a normalized root mean squared error of 0.07, and a mean absolute percentage error of 13.7%. This study demonstrates that aerial imagery with machine learning techniques can be a reliable, efficient, and effective tool for pre-harvest cotton yield prediction.

Adaptive Anisotropic Morphological Filtering Based on Co-Circularity of Local Orientations

The algorithm presented in this paper is an application of a general framework for morphological processing of signals on weighted graphs. Here we apply it to images by defining what we call a co-circularity graph. In this graph, the vertices are the pixels and the weighted edges depend on a consistency criterion (co-circularity) between local orientations estimated from the structure tensors. This graph induces anisotropic adaptive morphological operators which are related both to anisotropic diffusion in images and path optimality in graphs. We present several applications such as the enhancement of fiber-like structures, completion of interrupted edges and the regularization of grayscale images. We also discuss the parameters setting depending on the application.

Morphological geodesic active contour algorithm for the segmentation of the histogram‐equalized welding bead image edges

Assessment and evaluation are the essential processes of industrially manufactured products for the determination of the quality and quantity of products. They give justifications in a practical way about whether the machine is perfect or imperfect, which can lead to a better or poorer production. In this study, the authors propose an algorithm that uses morphological geodesic active contour and image processing techniques to perform segmentation and assess the performance of a robot used to manufacture welding beads. The algorithm has four parameters which are pre-processed images, balloon force, smoothing parameter, and number of iterations. To pre-process the images, the algorithm uses an inverse Gaussian gradient operator for edge detection and applies the histogram equalization method to level the distribution. To detect the external contour of the bead, the level set is initialized as the region of interest whereby a balloon force can inflate or deflate towards the edges. To smoothen the contour, a smoothing parameter is applied to convert the jagged lines into a curve over a reasonable number of iterations. Based on the experimental results, the authors’ algorithm used a fixed balloon force of −2, a smoothing parameter value of 4, and 40 iterations to segment images obtained from three different environments. The computation time for the segmentation and evaluation of one image was 0.70, 0.61, and 0.67 s for datasets with high brightness, low brightness, and normal brightness, respectively. Additionally, the authors’ proposed algorithm achieved an outstanding performance of 0.9954, 0.9843, 0.9892, and 0.9435 in terms of recall, precision, F-measure, and IOU, respectively. To justify the performance of the authors’ proposed algorithm, the authors compared it with the existing algorithms and found that it worked better than all the others for segmentation, although it lagged behind the entropy-based algorithm in terms of speed.