Texture Image Research Articles

An Experimental Investigation into the Enhancement of Surface Quality of Inconel 718 Through Axial Ultrasonic Vibration-Assisted Grinding in Dry and MQL Environments

Ultrasonic vibration-assisted grinding (UVAG) has proven to be beneficial for grinding difficult-to-machine materials. This work attempts to enhance the grinding performance of Inconel 718 through a comprehensive study of UVAG characteristics. Grinding experiments were performed in both dry and Minimum Quantity Lubrication (MQL) environments, and assessment of the grinding forces, specific energy, residual stress, and surface topography was done. A substantial reduction of both surface roughness and grinding force components was observed in UVAG compared to conventional grinding (CG). Utilizing UVAG with MQL at the maximum vibration amplitude led to a 64% reduction in tangential grinding force and a 51% decrease in roughness parameter, Ra, when compared to CG conducted in a dry environment. The high-frequency indentations of the abrasives in UVAG generated compressive residual stresses on the ground surface. Surface parameters pointed to uniform texture and SEM images showed widening of abrasive grain tracks on the workpiece surface during UVAG. The utilization of UVAG under MQL produced a synergistic impact and resulted in the lowest grinding forces, specific energy, and optimal surface quality among all the grinding conditions investigated. Overall analysis of the results indicated that the axial configuration of the vibration set-up is favorable for UVAG, and the high-frequency periodic separation-cutting characteristic of the process improves lubricating efficiency and grinding performance.

Distinguishing Structures from Textures by Patch‐based Contrasts around Pixels for High‐quality and Efficient Texture filtering

AbstractIt is still challenging with existing methods to distinguish structures from texture details, and so preventing texture filtering. Considering that the textures on both sides of a structural edge always differ much from each other in appearances, we determine whether a pixel is on a structure edge by exploiting the appearance contrast between patches around the pixel, and further propose an efficient implementation method. We demonstrate that our proposed method is more effective than existing methods to distinguish structures from texture details, and our required patches for texture measurement can be smaller than the used patches in existing methods by at least half. Thus, we can improve texture filtering on both quality and efficiency, as shown by the experimental results, e.g., we can handle the textured images with a resolution of 800 × 600 pixels in real‐time. (The code is available at https://github.com/hefengxiyulu/MLPC)

Lossless Recompression of Vector Quantization Index Table for Texture Images Based on Adaptive Huffman Coding Through Multi-Type Processing

With the development of the information age, all walks of life are inseparable from the internet. Every day, huge amounts of data are transmitted and stored on the internet. Therefore, to improve transmission efficiency and reduce storage occupancy, compression technology is becoming increasingly important. Based on different application scenarios, it is divided into lossless data compression and lossy data compression, which allows a certain degree of compression. Vector quantization (VQ) is a widely used lossy compression technology. Building upon VQ compression technology, we propose a lossless compression scheme for the VQ index table. In other words, our work aims to recompress VQ compression technology and restore it to the VQ compression carrier without loss. It is worth noting that our method specifically targets texture images. By leveraging the spatial symmetry inherent in these images, our approach generates high-frequency symbols through difference calculations, which facilitates the use of adaptive Huffman coding for efficient compression. Experimental results show that our scheme has better compression performance than other schemes.

PBNet: Combining Transformer and CNN in Passport Background Texture Printing Image Classification

Passport background texture classification has always been an important task in border checks. Current manual methods struggle to achieve satisfactory results in terms of consistency and stability for weakly textured background images. For this reason, this study designs and develops a CNN and Transformer complementary network (PBNet) for passport background texture image classification. We first design two encoders by Transformer and CNN to produce complementary features in the Transformer and CNN domains, respectively. Then, we cross-wisely concatenate these complementary features to propose a feature enhancement module (FEM) for effectively blending them. In addition, we introduce focal loss to relieve the overfitting problem caused by data imbalance. Experimental results show that our PBNet significantly surpasses the state-of-the-art image segmentation models based on CNNs, Transformers, and even Transformer and CNN combined models designed for passport background texture image classification.

Advancements in smart agriculture through innovative weed management using wavelet-based convolution neural network

Smart agriculture has shifted the paradigm by integrating advanced technologies, particularly weed management. This paper introduces an innovative approach to weed control by applying a Wavelet-based Convolution Neural Network (WCNN). In the era of precision agriculture, our study explores the integration of WCNN into real-world scenarios, emphasizing its adaptability to diverse environmental conditions. Utilizing the spatial-frequency analysis features of wavelets and convolutional neural networks, the WCNN model is the most effective at finding weeds, classifying them, and managing them specifically in agricultural fields in real-time. This research contributes to the scientific discourse on smart agriculture and addresses the challenges of invasive weeds, presenting a sustainable solution for optimizing resource utilization. Our investigation includes a detailed exploration of WCNN’s adaptive learning mechanisms and dynamic adjustment to changing agricultural landscapes. The model seamlessly integrates with existing smart farming infrastructure, showcasing a substantial reduction in manual intervention and a simultaneous increase in agricultural productivity. We incorporate fog computing and resource optimization into our framework, enhancing the efficiency of onboard data processing. To evaluate the real-world efficacy of WCNN, we conducted comprehensive experiments in texture classification and image labelling using two distinct datasets: the plant seedling and soybean weed datasets. Results demonstrate the superior performance of WCNN, achieving higher accuracy in training and test scenarios with significantly fewer parameters than traditional CNNs. For the soybean weed dataset, WCNN achieved remarkable accuracy in the training (0.9970) and testing (0.9987) phases, with correspondingly low losses of 0.0109 and 0.0048. The WCNN model demonstrated high accuracy during training (0.9739) and testing (0.9902), with minimal losses of 0.0898 and 0.0239 in the plant seedling dataset.

Single-use commercial bio-based plastics under environmental degradation conditions: Is their biodegradability and compostability a fact?

Evaluating compostability is increasingly essential for proving commercial bio-based cutlery or packaging since these materials must biodegrade under controlled conditions quickly. Utensils for eating represent Mexico's most popular consumer single-use materials, and Mexican regulations based on biodegradation or compostability are still vague and lack scientific evaluations. This study analyzed three bio-based polymeric materials (bags, dishes, and forks) from commercial brands following Mexican regulations and using various analytical techniques to verify their biodegradability and compostability. First, weight loss measurements, stress-strain tests, and topographic imaging were applied for preliminary observations at the macro scale up to 90 days of compostability. Besides, spectroscopy, microscopy, and thermal techniques indicate changes and behavior of the bio-based materials depending on the composition. The results suggest that bags exhibited the highest decomposition rate (80 %) compared to dishes and forks. Similarly, mechanical resistance indicates a reduction of 62 % for bags, 30 % for dishes, and almost none for forks. Texture image analysis revealed that the complexity and roughness of the materials increased over time, correlating with the physical changes observed. These results indicate minimal surface topography changes and higher stiffness for dishes and forks, indicating low biodegradability. SEM images supported these findings, showing surface degradation in bags and dishes but not in forks. FTIR and XRD analyses confirmed the presence of polyamide (bags) and polypropylene (dishes and forks). These results reduce biodegradation and differ from the claims made by manufacturers. The thermal analysis found similar results, indicating that the materials' thermal stability decreased after degradation, which is related to lower biodegradability and compostability. Overall, the study concluded only bags meet the criteria for compostability in national regulations. However, dishes and forks made of petroleum-derived polymers have higher resistance to natural and microbial degradation.

Classification of Kudoa thyrsites infected and uninfected fish using a handheld near-infrared spectrophotometer, SIMCA and PLS-DA.

Kudoa thyrsites infection of marine fish typically results in myoliquefaction, which is only apparent 24 to 56 h post-mortem. The traditional methods for the detection of K. thyrsites infected fish are time-consuming and destructive, reducing its marketability. This poses a challenge for the fish industry to remove infected fish before it reaches the market or further processing activities. This study investigated the use of near-infrared (NIR) spectroscopy, in combination with soft independent modelling of class analogy (SIMCA) and partial least square discriminant analysis (PLS-DA), for discriminating K. thyrsites infected fish from uninfected fish. Performance of the classification models was evaluated by calculating the sensitivity, specificity and precision. A total of 334 fish samples (200 sardine, 64 hake and 70 kingklip) were used for this study. Infection of K. thyrsites was determined with the use of qPCR assays. Ninety per cent (90%) of the sardine samples, 78% of the hake samples and 37% of the kingklip samples were infected. Class groups of infected and uninfected fish samples were created for the purpose of generating SIMCA and PLS-DA classification models for each species of fish, as well as for a species independent data set. Principal component analysis (PCA) of NIR spectra did not show any clustering for infected and uninfected samples. Calibration and test sample sets were generated for the purpose of building and testing the SIMCA and PLD-DA classification models. SIMCA and PLS-DA were unable to classify test samples correctly into the two classes. The number of misclassifications (NMC) was higher for the SIMCA models than for the PLS-DA models, with more than 60% incorrectly classified. SIMCA classified most of the test samples into both classes. The precision for PLS-DA were 89% for sardine, 81% for hake, 0% for kingklip and 87% for species independent models, however, most samples were classified at infected. The use of NIR spectroscopy and classification models such as SIMCA and PLS-DA showed limited use as a method to distinguish between K. thyrsites infected and uninfected fish samples. Textural and chemical changes during extended frozen storage of the fish samples may have masked the effects associated with K. thyrsites infection. Further studies are suggested where NIR spectroscopy is used in combination with texture analysis and image spectroscopy.

Learning graph-Fourier spectra of textured surface images for defect localization

In the realm of industrial manufacturing, product inspection remains a significant bottleneck, with only a small fraction of manufactured items undergoing inspection for surface defects. Advances in imaging systems and AI can allow automated full inspection of manufactured surfaces. However, even the most contemporary imaging and machine learning methods perform poorly for detecting defects in images with highly textured backgrounds, that stem from diverse manufacturing processes. This paper introduces an approach based on graph Fourier analysis to automatically identify defective images, as well as crucial graph Fourier coefficients that inform the defects in the images. The approach thereby facilitates precise localization and characterization of defects, amidst highly textured backgrounds. A convolutional neural network model (1D-CNN) was trained with the coefficients of the graph Fourier transform of the images as the input to identify, with classification accuracy of 99.4%, if the image contains a defect. An explainable AI method using SHAP (SHapley Additive exPlanations) was used to further analyze the trained 1D-CNN model to discern important spectral coefficients for each image. This approach sheds light on the crucial contribution of low-frequency graph eigen waveforms to precisely localize surface defects in images, thereby advancing the realization of zero-defect manufacturing.

Multimodal medical image fusion based on interval gradients and convolutional neural networks

Many image fusion methods have been proposed to leverage the advantages of functional and anatomical images while compensating for their shortcomings. These methods integrate functional and anatomical images while presenting physiological and metabolic organ information, making their diagnostic efficiency far greater than that of single-modal images. Currently, most existing multimodal medical imaging fusion methods are based on multiscale transformation, which involves obtaining pyramid features through multiscale transformation. Low-resolution images are used to analyse approximate image features, and high-resolution images are used to analyse detailed image features. Different fusion rules are applied to achieve feature fusion at different scales. Although these fusion methods based on multiscale transformation can effectively achieve multimodal medical image fusion, much detailed information is lost during multiscale and inverse transformation, resulting in blurred edges and a loss of detail in the fusion images. A multimodal medical image fusion method based on interval gradients and convolutional neural networks is proposed to overcome this problem. First, this method uses interval gradients for image decomposition to obtain structure and texture images. Second, deep neural networks are used to extract perception images. Three methods are used to fuse structure, texture, and perception images. Last, the images are combined to obtain the final fusion image after colour transformation. Compared with the reference algorithms, the proposed method performs better in multiple objective indicators of QEN\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$Q_{EN}$$\\end{document}, QNIQE\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$Q_{NIQE}$$\\end{document}, QSD\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$Q_{SD}$$\\end{document}, QSSEQ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$Q_{SSEQ}$$\\end{document} and QTMQI\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$Q_{TMQI}$$\\end{document}.

Computation Legendre moments using image analysis technique

Abstract In this paper, phase transition temperatures of nematic liquid crystals: 4-n alkoxy benzoic acid (nOBA, where n = 4 and 6) are determined from the Legendre moments of liquid crystal textures. Here, the image analysis technique in conjunction with polarizing optical microscope is used to compute the Legendre moments from the textures of liquid crystals. As a function of temperature, Liquid crystal textures are recorded from the arthroscopic mode of optical microscope. Changes in textural features of the sample with respect to temperature bring the variations in intensity values of liquid crystal textures. The brightness image that is generated from the liquid crystal textural patterns separates the information about fluctuations in intensity. In this approach, Legendre moments are computed from the brightness images of the liquid crystal textures and are plotted as a function of temperature. Significant variations in the brightness Legendre moment curve provide insight into the liquid crystal transition temperatures. Results obtained from the present methodology are compared with gray color textures data and the experimental technique of differential scanning calorimetry and are in good agreement with each other. This method is an objective and quantitative technique whereas POM method is qualitative and subjective.

Optical directional differential operation enabled visual chirality detection.

Directional differential operation can extract the changes of directional information from complex signals, and plays an important role in target recognition and texture image processing. Here, we propose an optical directional differential operation based on large cross-polarization rotation, and realize the visual detection of chiral enantiomers. By using cross-polarization rotation in a specified direction, we design a corresponding directional spatial spectral transfer function whose transmission efficiency increases as the incident angle approaches the Brewster angle. The differential direction can be adjusted by changing the initial polarization state, and can be used to detect the concentration of chiral solutions. Finally, we apply the directional differential operation to achieve the visual detection of chiral enantiomers.

Image inpainting algorithm based on inference attention module and two-stage network

Current image inpainting techniques often yield distorted structures or blurred textures that clash with the surrounding context, particularly when addressing extensive missing regions or highly textured images. These methods often struggle to reconstruct realistic and coherent image structures. To address this challenge, we introduce a two-stage network image restoration approach that leverages an inferential attention mechanism. In the first stage, an edge generation network is employed to produce plausible phantom edge information. Subsequently, an image complementation network is utilized to complete the image restoration process. To enhance the visual realism and restoration accuracy of the generated images, we incorporate an inferential attention mechanism within the image complementation network. This mechanism effectively mitigates inconsistencies in the generated features, leading to the production of more meaningful and effective information. We evaluate our proposed method on benchmark datasets, including Places2 dataset, CelebFaces Attributes dataset, and Paris StreetView dataset. The experimental results can demonstrate that the proposed method achieves better Structural Similarity Index and Peak Signal-to-noise Ratio than others. These metrics indicate superior performance compared to existing image inpainting methods, delivering higher image inpainting accuracy and more realistic image reconstruction effect.

Neuronal and Behavioral Responses to Naturalistic Texture Images in Macaque Monkeys.

The visual world is richly adorned with texture, which can serve to delineate important elements of natural scenes. In anesthetized macaque monkeys, selectivity for the statistical features of natural texture is weak in V1, but substantial in V2, suggesting that neuronal activity in V2 might directly support texture perception. To test this, we investigated the relation between single cell activity in macaque V1 and V2 and simultaneously measured behavioral judgments of texture. We generated stimuli along a continuum between naturalistic texture and phase-randomized noise and trained two macaque monkeys to judge whether a sample texture more closely resembled one or the other extreme. Analysis of responses revealed that individual V1 and V2 neurons carried much less information about texture naturalness than behavioral reports. However, the sensitivity of V2 neurons, especially those preferring naturalistic textures, was significantly closer to that of behavior compared with V1. The firing of both V1 and V2 neurons predicted perceptual choices in response to repeated presentations of the same ambiguous stimulus in one monkey, despite low individual neural sensitivity. However, neither population predicted choice in the second monkey. We conclude that neural responses supporting texture perception likely continue to develop downstream of V2. Further, combined with neural data recorded while the same two monkeys performed an orientation discrimination task, our results demonstrate that choice-correlated neural activity in early sensory cortex is unstable across observers and tasks, untethered from neuronal sensitivity, and therefore unlikely to directly reflect the formation of perceptual decisions.

Dual cross perception network with texture and boundary guidance for camouflaged object detection

Camouflaged object detection (COD) is a task needs to segment objects that subtly blend into their surroundings effectively. Edge and texture information of the objects can be utilized to reveal the edges of camouflaged objects and detect texture differences between camouflaged objects and the surrounding environment. However, existing methods often fail to fully exploit the advantages of these two types of information. Considering this, our paper proposes an innovative Dual Cross Perception Network (DCPNet) with texture and boundary guidance for camouflaged object detection. DCPNet consists of two essential modules, namely Dual Cross Fusion Module (DCFM) and the Subgroup Aggregation Module (SAM). DCFM utilizes attention techniques to emphasize the information that exists in edges and textures by cross-fusing features of the edge, texture, and basic RGB image, which strengthens the ability to capture edge information and texture details in image analysis. SAM gives varied weights to low-level and high-level features in order to enhance the comprehension of objects and scenes of various sizes. Several experiments have demonstrated that DCPNet outperforms 13 state-of-the-art methods on four widely used assessment metrics.

Real-Space Imaging of Intrinsic Symmetry-Breaking Spin Textures in a Kagome Lattice.

The electronic orders in kagome materials have emerged as a fertile platform for studying exotic quantum states, and their intertwining with the unique kagome lattice geometry remains elusive. While various unconventional charge orders with broken symmetry is observed, the influence of kagome symmetry on magnetic order has so far not been directly observed. Here, using a high-resolution magnetic force microscopy, it is, for the first time, observed a new lattice form of noncollinear spin textures in the kagome ferromagnet in zero magnetic field. Under the influence of the sixfold rotational symmetry of the kagome lattice, the spin textures are hexagonal in shape and can further form a honeycomb lattice structure. Subsequent thermal cycling measurements reveal that these spin textures transform into a non-uniform in-plane ferromagnetic ground state at low temperatures and can fully rebuild at elevated temperatures, showing a strong second-order phase transition feature. Moreover, some out-of-plane magnetic moments persist at low temperatures, supporting the Kane-Mele scenario in explaining the emergence of the Dirac gap. The observations establish that the electronic properties, including both charge and spin orders, are strongly coupled with the kagome lattices.

Soft X-Ray Phase Nanomicroscopy of Micrometer-Thick Magnets

Imaging of nanoscale magnetic textures within extended material systems is of critical importance to both fundamental research and technological applications. While high-resolution magnetic imaging of thin nanoscale samples is well established with electron and soft x-ray microscopy, the extension to micrometer-thick systems currently requires hard x rays, which limits high-resolution imaging to rare-earth magnets. Here, we overcome this limitation by establishing soft x-ray magnetic imaging of micrometer-thick systems using the pre-edge phase x-ray magnetic circular dichroism signal, thus making possible the study of a wide range of magnetic materials. By performing dichroic spectroptychography, we demonstrate high spatial resolution imaging of magnetic samples up to 1.7 μm thick, an order of magnitude higher than conventionally possible with soft x-ray absorption-based techniques. We demonstrate the applicability of the technique by harnessing the pre-edge phase to image thick chiral helimagnets, and naturally occurring magnetite particles, gaining insight into their three-dimensional magnetic configuration. This new regime of magnetic imaging makes possible the study of extended non-rare-earth systems that have until now been inaccessible, including magnetic textures for future spintronic applications, non-rare-earth permanent magnets for energy harvesting, and the magnetic configuration of giant magnetofossils. Published by the American Physical Society 2024

Indoor thermal environment simulation of jade carving design process based on light sensing texture image processing

As a traditional art form, jade carving has gradually combined with modern technology in recent years, introducing new technologies in the design and production process. Indoor thermal environment has an important influence on the creation process of jade carving, especially the influence of temperature and humidity on the physical characteristics of jade carving materials and its processing. In this study, the influence of indoor thermal environment on jade carving design process was simulated by light sensing texture image processing technology, so as to optimize the design scheme and ensure the quality and artistic expression of jade carving works. The temperature, humidity and illumination data of indoor environment are obtained by light sensing technology, and the thermal environment required in jade carving design is simulated by texture image processing method. The thermal environment model was constructed to analyze the performance of jade carving materials under different conditions and evaluate the possible changes in the processing process. The simulation results show that the temperature and humidity of indoor thermal environment have significant effects on the toughness, hardness and detail performance of jade carving materials. At temperatures above 25° C and relatively low humidity, jade carving materials show better cutting results and surface finish. Texture image processing technology can effectively capture and analyze the subtle changes of materials, which provides a reliable basis for design decisions.

Robust ROI Detection in Whole Slide Images Guided by Pathologists' Viewing Patterns.

Deep learning techniques offer improvements in computer-aided diagnosis systems. However, acquiring image domain annotations is challenging due to the knowledge and commitment required of expert pathologists. Pathologists often identify regions in whole slide images with diagnostic relevance rather than examining the entire slide, with a positive correlation between the time spent on these critical image regions and diagnostic accuracy. In this paper, a heatmap is generated to represent pathologists' viewing patterns during diagnosis and used to guide a deep learning architecture during training. The proposed system outperforms traditional approaches based on color and texture image characteristics, integrating pathologists' domain expertise to enhance region of interest detection without needing individual case annotations. Evaluating our best model, a U-Net model with a pre-trained ResNet-18 encoder, on a skin biopsy whole slide image dataset for melanoma diagnosis, shows its potential in detecting regions of interest, surpassing conventional methods with an increase of 20%, 11%, 22%, and 12% in precision, recall, F1-score, and Intersection over Union, respectively. In a clinical evaluation, three dermatopathologists agreed on the model's effectiveness in replicating pathologists' diagnostic viewing behavior and accurately identifying critical regions. Finally, our study demonstrates that incorporating heatmaps as supplementary signals can enhance the performance of computer-aided diagnosis systems. Without the availability of eye tracking data, identifying precise focus areas is challenging, but our approach shows promise in assisting pathologists in improving diagnostic accuracy and efficiency, streamlining annotation processes, and aiding the training of new pathologists.

Intelligent Fault Diagnosis Method for Rotating Machinery Based on Recurrence Binary Plot and DSD-CNN.

To tackle the issue of the traditional intelligent diagnostic algorithm's insufficient utilization of correlation characteristics within the time series of fault signals and to meet the challenges of accuracy and computational complexity in rotating machinery fault diagnosis, a novel approach based on a recurrence binary plot (RBP) and a lightweight, deep, separable, dilated convolutional neural network (DSD-CNN) is proposed. Firstly, a recursive encoding method is used to convert the fault vibration signals of rotating machinery into two-dimensional texture images, extracting feature information from the internal structure of the fault signals as the input for the model. Subsequently, leveraging the excellent feature extraction capabilities of a lightweight convolutional neural network embedded with attention modules, the fault diagnosis of rotating machinery is carried out. The experimental results using different datasets demonstrate that the proposed model achieves excellent diagnostic accuracy and computational efficiency. Additionally, compared with other representative fault diagnosis methods, this model shows better anti-noise performance under different noise test data, and it provides a reliable and efficient reference solution for rotating machinery fault-classification tasks.

Surface texture image classification of carbon/phenolic composites in extreme environments using deep learning

AbstractThe classification of ablation images holds significant practical value in thermal protection structures, as it enables the assessment of heat and corrosion resistance of composites. This paper proposes an image‐based deep learning framework to identify the surface texture of carbon/phenolic composites ablative images. First, ablation experiments and collection of surface texture images of carbon/phenolic composites under different thermal environments were conducted in an electric arc wind tunnel. Then, a deep learning model based on a convolutional neural network (CNN) is developed for ablative image classification. The pre‐trained network is ultimately employed as the input for transfer learning. The network's feature extraction layer is trained using the ImageNet dataset, while the global average pooling addresses specific classification tasks. The test results demonstrate that the proposed method effectively classifies the relatively small surface texture dataset, enhances the classification performance of ablative surface texture with an accuracy of up to 97.6%, and exhibits robustness and generalization capabilities.Highlights The paper proposes a new deep learning classification method for ablative images. A model highly sensitive to small and weak features is built. Transfer learning and data enhancement techniques are introduced into classification.