Morphological Processing Research Articles

Two-stage container keyhole location algorithm based on optimized SSD and adaptive threshold

The automation transformation of container lifting operations is one of the main technical issues in Rail-Truck intermodal transportation. To solve this problem, this paper analyzes the advantages and disadvantages of several existing container positioning methods, and proposed a container corner holes location detection method based on lightweight convolutional neural network and adaptive morphological image processing algorithm. This method locates the container through a visual sensor that shoots from top to bottom. In order to improve the positioning accuracy and calculation speed while maintaining a high recognition rate, this method first uses a lightweight SSD detector to quickly detect the rough coordinates of the container corner hole in the image. On this basis, the smallest rectangle detection method based on adaptive HSV filtering is used to detect the precise coordinates of the corner hole in the image. Experiments show that the recognition rate of this algorithm in the initial positioning process reaches 94.2%, the recognition rate of secondary positioning reaches 87.4%, and the final positioning error is lower than 3.765 pixels and 2.81 pixels in the heading and lateral directions, respectively. The overall calculation time of the algorithm can realize the positioning calculation of 30 frames per second, which shows that this method takes into account the characteristics of high measurement accuracy and high calculation speed on the basis of high recognition rate.

Effect of V–III Ratio‐Based Growth Mode on the Surface Morphology, Strain Relaxation, and Dislocation Density of AlN Films Grown by Metal‐Organic Chemical Vapor Deposition

Herein, regulation of the growth mode through manipulating the V–III ratio during metal‐organic chemical vapor deposition (MOCVD) growth of AlN films on c‐plane sapphire substrates and optimization of the crystalline quality and surface morphology of AlN films are demonstrated. Based on detailed analyses of the evolution process of surface morphology, dislocation annihilation, and stress state of the AlN film during MOCVD growth, it is found that 2D rather than 3D growth mode introduces tensile stress as well as promotes edge‐type threading dislocation (ETD) propagation and crack generation, while 3D growth mode produces rough surface. These results would help to further understand the AlN growth mechanism and provide important guidance on the preparation of AlN‐related epitaxial structure.

Mass transfer approaches for CO2 separation in non-isothermal and non-adiabatic pressure swing adsorption system for biomethane upgrading

The mass transfer of CO2 over solid materials can be quite complex. The complexity arises as a result of the adsorbent’s complex surface morphology and dynamic process operating conditions. In this study, a new model was proposed to predict the mass transfer behavior of CO2 over silica gel in biogas upgrading process in which three mass transfer systems (lumped, micro-/macropore, and particle) were introduced. An overview for such mathematical models for CO2 mass transfer was conducted and then a new non-adiabatic and non-isothermal model for pressure swing adsorption (PSA) that is different from the existing models in the literature, i.e., adiabatic and isothermal models, was built and assessed. The new model was executed using Aspen Adsorption and validated using experimental data reported previously over silica gel bed. The validation was performed by comparing molar fractions of CH4 and CO2, temperature profile of silica gel bed, and bio-CH4 purity and recovery with previously reported experimental data. The particle mass transfer system, which estimated the adsorption rate and gas concentration profile in the adsorbent’s particle, recorded the highest bio-CH4 purity of 99.12%. Lumped approach, which comprised mass transfer resistances in one overall resistance and ignored the dispersion effect, recorded bio-CH4 purity of 97.63% followed by micro-/macropore, which accounted micro and macropore resistance individually, with purity of 84.65%. Particle mass transfer approach yielded the highest bio-CH4 recovery of 96.745%, followed by lumped (95.41077%), and micro/macropore (82.433%) systems. Particle mass transfer was one of the most realistic systems in predicting the CO2 diffusion over silica gel due to its comprehensive heat/mass transfer consideration along with its assumption of uniform pore structure that fitted well with the true nature of the microporous silica character.

Rapid detection for the surface and hidden crack of solid propellant motor charge using laser arrays scanning thermography (LAsST)

Laser arrays scanning thermography (LAsST) was introduced to detect the surface and hidden cracks of solid propellant motor charge. A series of simulated surface cracks of solid propellant charge was prepared and detected by LAsST. The thermal-wave response signal is reconstructed by truncation pseudo-static matrix reconstruction (TC-PSMR), and then, the thermal image sequence similar to the impulse response is reconstructed. The principle of characteristic extraction algorithms [such as Fast Fourier transform (FFT), principal component analysis (PCA), and partial least squares regression (PLSR)] are introduced and employed to obtain the feature images from the thermal-wave response. The morphological matrix is used to denoise the processed image. The experimental results demonstrated that the defects of the PLSR image and FFT phase image are more obvious, and the image noise after morphological matrix processing is significantly reduced. Meanwhile, the LAsST approach can enable rapid detection for the surface and hidden cracks of solid propellant motor charge.

Enhanced photocatalytic hydrogen evolution under visible-light using C, N co-doped mesoporous TiO2 nanocrystals templated by ionic liquids

Introducing heteroatoms into TiO2 photocatalysts is an effective way to improve its visible-light driven photocatalytic performance. However, the strategies to realize it always require hazardous reagents and/or high temperature. Herein, we put forward a facile and environmentally benign strategy for synthesizing C, N co-doped mesoporous TiO2 nanocrystals. The preparation consists of microwave heating of a simple mixture of H2O, imidazolium-based ionic liquid (IL) and tetrabutyl titanate and subsequent heat-treatment. During the synthesis, the IL acts as reaction solvent, microwave absorbent, C and N sources, and structure-directing agent. Such a strategy not only simplifies the synthesis by avoiding the employment of additional reactants and structure-directing agents towards constructing TiO2 with small sizes, well-developed mesoporous structure and abundant C, N dopants in the lattices, but also lows down the post-processing temperature, which can relieve the secondary growth of TiO2 grains. Benefiting from the mesoporous structure and heteroatom co-doping, the synthesized material shows high specific-surface area, excellent hydrophilicity, elevated valence-band edges, abundant defect levels and narrowed band-gap, which contribute to the improved visible-light absorption, reduced photogenerated electron-hole recombination, and rapid charge transfer and surface-catalyzed reactions. As a result, the material exhibits higher photocatalytic H2 generation performance under the visible-light irradiation than many conventional types of doped TiO2. Besides, the formation mechanism of this TiO2 and the relationships of IL with the morphology and co-doping process have been revealed. This work highlights great potential of IL-assisted microwave synthesis in developing heteroatoms co-doped TiO2 with enhanced photoactivity and application in the solar energy conversion.

Comparative analysis of deep learning classifiers for diabetic retinopathy identification and detection

ABSTRACT Diabetic retinopathy (DR) is a micro vascular problem caused by diabetes that can lead to loss of sight. The early detection of diabetic retinopathy is important to avoid the severity of sightlessness. In this manuscript, a comparative analysis of several deep learning methods for DR identification is proposed. The input fundus images are taken from a standard dataset pre-processed by the Mathematical Morphology process. Moreover, the images are segregated using a Multilevel segmentation of the Region of interest (ROI) based on the split and merge algorithm. After that, an original deep learning architecture is utilized to categorize the segregated fundus images. Deep learning methods, such as Convolution neural network (CNN), Recurrent Neural Network (RNN), Support Vector Machine (SVM), Fuzzy K-means cluster (FKM) and Discriminant Analysis (DA) are proposed to classify the DR. The proposed DR identification and detection with CNN provides 65.54% SP, 100% SE, 78.54% SV and 96.95% ACC. Finally, CNN shows better performance than other classifiers.

Explainable multiple abnormality classification of chest CT volumes

Understanding model predictions is critical in healthcare, to facilitate rapid verification of model correctness and to guard against use of models that exploit confounding variables. We introduce the challenging new task of explainable multiple abnormality classification in volumetric medical images, in which a model must indicate the regions used to predict each abnormality. To solve this task, we propose a multiple instance learning convolutional neural network, AxialNet, that allows identification of top slices for each abnormality. Next we incorporate HiResCAM, an attention mechanism, to identify sub-slice regions. We prove that for AxialNet, HiResCAM explanations are guaranteed to reflect the locations the model used, unlike Grad-CAM which sometimes highlights irrelevant locations. Armed with a model that produces faithful explanations, we then aim to improve the model’s learning through a novel mask loss that leverages HiResCAM and 3D allowed regions to encourage the model to predict abnormalities based only on the organs in which those abnormalities appear. The 3D allowed regions are obtained automatically through a new approach, PARTITION, that combines location information extracted from radiology reports with organ segmentation maps obtained through morphological image processing. Overall, we propose the first model for explainable multi-abnormality prediction in volumetric medical images, and then use the mask loss to achieve a 33% improvement in organ localization of multiple abnormalities in the RAD-ChestCT dataset of 36,316 scans, representing the state of the art. This work advances the clinical applicability of multiple abnormality modeling in chest CT volumes.

Study on micro-energy consumption model of ultrafine grinding coal particles

Based on the particle size distribution and surface physicochemical characteristics of coal particles, a micro-energy consumption model of coal particles in the ultrafine grinding process was proposed. This model explained the microscopic mechanical behavior of the ultrafine grinding process by defining factors such as agglomeration between particles and the change in particle surface properties as the microscopic constraint energy of particles. Under the multi-domain fractal law of particle size distribution, the truncated particle size was introduced to eliminate the mutual interference between size grades, and the expression of the micro-energy consumption law in the ultrafine grinding process of coal particles was derived by combining with the surface area theory of macroscopic crushing. Through scanning electron microscopy and infrared spectroscopy analysis, it was confirmed that the evolutionary process of coal particle surface morphology and the fracture law of chemical bonds were in line with the viewpoint of the micro-energy consumption model.

Biocompatible Ionic Liquids in High-Performing Organic Electrochemical Transistors for Ion Detection and Electrophysiological Monitoring.

Organic electrochemical transistors (OECTs) have recently attracted attention due to their high transconductance and low operating voltage, which makes them ideal for a wide range of biosensing applications. Poly-3,4-ethylenedioxythiophene:poly-4-styrenesulfonate (PEDOT:PSS) is a typical material used as the active channel layer in OECTs. Pristine PEDOT:PSS has poor electrical conductivity, and additives are typically introduced to improve its conductivity and OECT performance. However, these additives are mostly either toxic or not proven to be biocompatible. Herein, a biocompatible ionic liquid [MTEOA][MeOSO3] is demonstrated to be an effective additive to enhance the performance of PEDOT:PSS-based OECTs. The influence of [MTEOA][MeOSO3] on the conductivity, morphology, and redox process of PEDOT:PSS is investigated. The PEDOT:PSS/[MTEOA][MeOSO3]-based OECT exhibits high transconductance (22.3 ± 4.5 mS μm-1), high μC* (the product of mobility μ and volumetric capacitance C*) (283.80 ± 29.66 F cm-1 V-1 s-1), fast response time (∼40.57 μs), and excellent switching cyclical stability. Next, the integration of sodium (Na+) and potassium (K+) ion-selective membranes with the OECTs is demonstrated, enabling selective ion detection in the physiological range. In addition, flexible OECTs are designed for electrocardiography (ECG) signal acquisition. These OECTs have shown robust performance against physical deformation and successfully recorded high-quality ECG signals.

Detection of infestation by striped stem‐borer (Chilo suppressalis) in rice based on hyperspectral imaging

AbstractStriped stem‐borer (SSB) larvae are one of the most serious insect pests that damages rice. In this study, the combined hyperspectral imaging and chemometric method are proposed detect the early SSB larvae infestation in rice stems. Firstly, hyperspectral reflectance data (390–1040 nm) of rice stems were collected for normal, 3 days of infection, 6 days of infection, and 9 days of infection, respectively. On the spectra, the average spectra of the region of interest of the rice stalk samples were extracted, and 19 characteristic wavelengths were screened by the successive projection algorithm (SPA). Morphological processing was performed on the pseudo‐color (generally known as RGB) images of rice stalk samples on the images. Then five image texture features based on the gray‐scale co‐occurrence matrix (GLCM) and nine image color features based on color moments were obtained. The evaluation models of SSB larval infestation based on full spectra, characteristic wavelengths, GLCM, color moments, and data fusion were developed using partial least squares discriminant analysis (PLS‐DA), respectively. The results show that performance of the PLS‐DA model based on the fusion of characteristic wavelengths and GLCM is best. The accuracy of both the comprehensive modeling set and the prediction set is above 97%. The accuracy of the degree of infestation classification is good for all categories, and only the samples on the third day of infection are slightly less accurate. In conclusion, the hyperspectral imaging technology can be used to effectively detect the infection degree situation of early SSB larval infestation.Practical ApplicationsChemical control using insecticides to manage SSB infestation on rice is still the farmers' first choice. However, in the middle and late stages of rice growth, SSB larvae may moth into the inside of the rice stalk, and the rice provides a shelter for SSB so that the sprayed pesticides cannot reach SSB larvae, which produce a large number of overwintering SSB. This leads to the misuse of chemical pesticides, which causes serious pollution to the environment, pesticide residues also cause serious harm to the human body, and it also makes the pest resistance to pesticides rapidly increase. Therefore, it is necessary to find a rapid and effective detection method to detect the early infestation of SSB larvae on rice to reduce the use of pesticides and improve rice yield.

Molecular dynamics analysis of the influence of ion implantation parameters on ultra-precision machining of silicon carbide

This paper aims to study the mechanism of ion implantation. Molecular dynamics software was used to simulate the process of 3C-SiC workpiece being ion implanted with different implant parameters. The influence of ion implantation parameters on workpiece damage, as well as on workpiece temperature, potential energy, dislocation defects, surface morphology, cutting force, stress and other processing parameters, is analyzed. The results show that with the increase of dose, the damage range inside the workpiece shows a periodic change, the potential energy and temperature generally show a rising trend, and the internal normal stress exhibits an increasing trend. In addition, the temperature, potential energy, cutting force and stress of the workpiece increase with the increasing inclination angle. Furthermore, the channel effect is obvious when the inclination angle of ion implantation is between 30° and 50°. Alternatively, for inclination angle greater than 70°, a scattering effect is distinctly observed. Moreover, the average cutting force decreases with the increase of injection distance. Lastly, the average temperature is lowest when the ion implantation region is in the middle of the workpiece, and the farther the implantation region is, the faster is the rise in temperature and potential energy.

The effect of oral vocabulary training on reading novel complex words.

Do readers benefit from their knowledge of the phonological form and meaning of stems when seeing them embedded in morphologically complex words for the first time in print? This question was addressed using a word learning paradigm. Participants were trained on novel spoken word stems and their meanings ("tump"). Following training, participants then saw the novel stems for the first time in print, either in combination with a real affix (tumpist, tumpor) or with a non-affix (tumpel, tumpain). Untrained items were also included to test whether the affix effect was modulated by the prior training of the spoken word stems. First, the complex words were embedded in meaningful sentences which participants read as their eye movements were recorded (first orthographic exposure). Second, participants were asked to read aloud and spell each individual complex novel word (second orthographic exposure). Participants spent less time fixating on words that included trained stems compared with untrained stems. However, the training effect did not change depending on whether stems were accompanied by a real affix or a non-affix. In the reading aloud and spelling tasks, there was no effect of training, suggesting that the effect of oral vocabulary training did not extend beyond the initial print exposure. The results indicate that familiarity with spoken stems influences how complex words containing those stems are processed when being read for the first time. Our findings highlight the flexibility and adaptability of the morphological processing system to novel complex words during the first print exposure.

Surface crack analysis of the steering shaft gear after carburizing and quenching

The phenomenon of surface cracks in steering shaft gear after carburizing and quenching was analyzed by combining experimental research, theoretical analysis, and numerical simulation. The chemical composition, crack morphology, non-metallic inclusions, microstructure, microhardness, and forging process were analyzed and studied by means of directly reading spectrometer, metallographic microscope, scanning electron microscope, energy dispersive spectrometer, and micro hardness tester. The evolution process of surface crack morphology was numerically simulated and reproduced with the help of Deform 3D point tracking technology. The results show that linear surface cracks are formed on the surface of the round steel because of Alumina-type non-metallic inclusions, which destroy the continuity of the steel matrix during the rolling process. Surface cracks become arc-shaped due to the regular plastic flow direction of the material and propagate in the depth direction during the forging process. Crack provide a channel for the diffusion of C atoms during carburizing and quenching process, therefore carburized layer with gradient distribution is formed on both sides of the crack. The carburized layer reacts with oxygen during quenching and subsequent tempering, resulting in a certain degree of oxidation on both sides of the crack and the crack tail. Therefore, the surface cracks found in the steering shaft gear after carburizing and quenching are originated from the rolling cracks formed by non-metallic inclusions of steel. The rolling cracks have undergone further evolution in the subsequent forging and carburizing quenching process.

Contribution of morphological awareness to reading fluency of children with and without dyslexia: evidence from a transparent orthography.

The goal of this study was to investigate the contribution of morphological awareness to reading fluency of children with and without dyslexia in a transparent orthography, such as the Greek one. The sample consisted of 256 Greek-speaking children (2nd grade: 32 dyslexic and 105 typical readers, 5th grade: 28 dyslexic and 91 typical readers). Morphological awareness was assessed with three tasks, examining inflectional, derivational, and compounding morphology. Reading fluency was evaluated at word, text, and silent level. The results indicated that dyslexic children both in 2nd and 5th grade face significant difficulties in inflectional, derivational, and compounding morphology compared to their peers. Hierarchical regression analyses revealed that morphological awareness significantly contributed to reading fluency of children with and without dyslexia, after controlling for non-verbal intelligence, vocabulary, and phonological awareness. Among typical readers, results indicated that inflectional, derivational, and compounding morphology had a small but significant effect on word, text, and silent reading fluency in 2ndgrade and derivational and inflectional morphology on text and silent reading fluency in 5th grade, after controlling for non-verbal intelligence, vocabulary, and phonological awareness. For dyslexic children, a moderate-to-large effect of inflectional and derivational morphology on text and word reading fluency was restricted to 2nd grade. Overall, morphological skills may play a supportive role in reading fluency of Greek children in first and last elementary grades. On the other hand, for Greek children facing reading problems morphological skills appeared to have a strong role in reading fluency only in first grades. Our study provided some preliminary data for the dyslexics' ability of morphological processing as a scaffolding skill for reading fluency. Implications of these findings for education are discussed.

Detection and localization of citrus fruit based on improved You Only Look Once v5s and binocular vision in the orchard.

Intelligent detection and localization of mature citrus fruits is a critical challenge in developing an automatic harvesting robot. Variable illumination conditions and different occlusion states are some of the essential issues that must be addressed for the accurate detection and localization of citrus in the orchard environment. In this paper, a novel method for the detection and localization of mature citrus using improved You Only Look Once (YOLO) v5s with binocular vision is proposed. First, a new loss function (polarity binary cross-entropy with logit loss) for YOLO v5s is designed to calculate the loss value of class probability and objectness score, so that a large penalty for false and missing detection is applied during the training process. Second, to recover the missing depth information caused by randomly overlapping background participants, Cr-Cb chromatic mapping, the Otsu thresholding algorithm, and morphological processing are successively used to extract the complete shape of the citrus, and the kriging method is applied to obtain the best linear unbiased estimator for the missing depth value. Finally, the citrus spatial position and posture information are obtained according to the camera imaging model and the geometric features of the citrus. The experimental results show that the recall rates of citrus detection under non-uniform illumination conditions, weak illumination, and well illumination are 99.55%, 98.47%, and 98.48%, respectively, approximately 2–9% higher than those of the original YOLO v5s network. The average error of the distance between the citrus fruit and the camera is 3.98 mm, and the average errors of the citrus diameters in the 3D direction are less than 2.75 mm. The average detection time per frame is 78.96 ms. The results indicate that our method can detect and localize citrus fruits in the complex environment of orchards with high accuracy and speed. Our dataset and codes are available at https://github.com/AshesBen/citrus-detection-localization.

Analisis Proses Morfologi pada Kolom Show dan Selebriti di Koran Harian Jawa Pos Edisi Juni 2022

This research discusses the morphology process that aims to identify forms of affixation, reduplication, and composition in the use of vocabulary in news texts in the Show & Celebrity column of the Jawa Pos Daily Newspaper June 2022 Edition. The approach in this study is to use a descriptive qualitative approach. Data collection is done by using observation techniques or observations that are observed through news texts and note-taking techniques. The results of this study include, (1) there are 18 data on the use of affixation, namely, prefix ber-, tar, {Me-}, and prefix {di-}, clofix {Me-}kan, clofix {Me-}i; (2) there are 2 data on the use of reduplication, namely songs and albums; and (3) there is 1 data on the use of composition, namely music videos.

Optical Investigation of the Influence of In-cylinder Flow and Mixture Inhomogeneity on Cyclic Variability in a Direct-Injection Spark Ignition Engine

In this study, a single-cylinder direct-injection spark-ignition research engine with full optical access was used to investigate the influence of the flow field and fuel/air mixing on cyclic variability, in particular in the early flame propagation. The engine was operated under lean-burn conditions at 1500 rpm. Two different injection strategies were compared, port-fuel injection (PFI) and direct injection (DI), the latter with early and late injection split about 2:1 in fuel mass. High-speed particle image velocimetry captured the flow in the tumble plane in the compression stroke. The velocity fields and the movement of the tumble vortex are analyzed. Simultaneously, a second camera detected the chemiluminescence of the flame, and the projected area of the line-of-sight-integrated flame luminosity was extracted through morphological image processing. By combining pressure-based combustion analysis and high-speed optical diagnostics, the early flame propagation and the flow field are correlated. In separate experiments the equivalence ratio was imaged for the DI at selected crank angles and correlated with CA10 to learn about the influence of mixture inhomogeneity on early flame propagation. With PFI, the flow near the spark plug just before ignition is closely related to the subsequent speed of combustion. The combustion-relevant flow features can be traced back in time to about –90 °CA. In contrast, the chosen DI scheme results in a decorrelation of flow and flame, and the equivalence ratio distribution at ignition becomes more important. For both flow and mixture fields, regions of high correlation with early-combustion metrics are typically associated with gradients in the multi-cycle average fields.

General and facile synthesis, formation process, and luminescence properties of well-defined rare earth oxides hollow spheres

Uniform and well-defined Ln2O3 (Ln = Gd and Lu) hollow spheres have been fabricated via a hard template directed method with 3-aminophenol-formaldehyde (AF) resin as a template. During the synthesis procedure, the core-shell structured AF@Ln(OH)CO3 precursor spheres are firstly obtained via a homogeneous precipitation method. The final Ln2O3 hollow spheres are synthesized by a subsequent calcination treatment, which inherit the uniformity, dispersity, and morphology of the precursors and exhibit a reduction in diameter. The crystal structure, morphology, and formation process of the hollow spherical products have been systematically studied. After incorporating the proper lanthanide activator ions such as Eu3+, Yb3+/Er3+, or Yb3+/Ho3+ ions into the Gd2O3 and Lu2O3 host spheres, the as-synthesized luminescent hollow spheres can exhibit the characteristic downshift and upconversion emissions from different activator ions. The LED devices fabricated by the Gd2O3:Eu3+ and Lu2O3:Eu3+ phosphors and LED chip can show dazzling red emission light, which directly proves that the as-obtained phosphors may have prospects in fields of optoelectronic devices. Furthermore, the facile and general strategy may pave a critical way for the preparation of the hollow spherical rare earth oxides with well-defined structure and eminent physicochemical properties.

Stray Light Nonuniform Background Elimination Method Based on Image Block Self-Adaptive Gray-Scale Morphology for Wide-Field Surveillance

Space-based wide-field surveillance systems are of great significance in maintaining the security of space resources by avoiding collisions between space targets. However, their performance is hindered by stray light phenomena. The nonuniform background noise caused by stray light significantly hampers subsequent target detection, leading to a high frequency of false alarms. To solve this problem, we propose a robust and accurate nonuniform background elimination method based on image block self-adaptive gray-scale morphology (IBSGM). First, we define two kinds of structural operators with different sizes and domains, which make full use of the difference between the target pixels and surrounding background pixels. Then, we block the original surveillance image and find the size of the largest target in each block by the minimum bounding rectangle method to determine the optimal size of the structural operator suitable for each block. Finally, we perform morphological processing using the defined structural operators to eliminate nonuniform backgrounds from images. Experimental results on simulated and real image datasets demonstrate that the proposed IBSGM method has higher precision in eliminating the nonuniform background when compared to other methods.

Influence mechanism of abrasive belt wear on fatigue resistance of TC17 grinding surface

Abrasive belt grains morphology and wear process have direct effect on grinding surface quality and service performance of titanium alloys. In order to increase fundamental understanding of abrasive belt wear on service performance of grinding surfaces, further provide targeted guidance for anti-fatigue surface grinding, this paper analyzed the surface characteristics and fatigue failure mechanism of the TC17 samples ground before and after abrasive belt wear. Results showed that for zirconium corundum abrasive belt with relatively independent abrasive grains, the fatigue failure of machined surface before abrasive belt wear is dominated by roughness. After zirconium corundum abrasive belt was worn, grinding surface roughness is reduced and surface mechanical properties are strengthened, which promotes the initiation of fatigue cracks from the subsurface and the improvement of fatigue life. Although the wear of electroplated diamond abrasive belt can reduce the surface roughness of grinding surface, the embedding and accumulation of abrasive grains hindered the chip removal and heat dissipation in material removal process, making the grinding surface easy to form heterogeneous ablative layer, and fatigue cracks initiated from the subsurface ablation area. The heterogeneous areas formed by surface burns will exacerbate surface stress concentration, which in turn will greatly reduce surface fatigue performance, it needs to be avoided as much as possible. Sharp electroplated diamond belts and worn zirconium corundum belts have excellent advantages in improving surface anti-fatigue performance.