Deep Learning Object Detection Research Articles

Determination of the Rate Dependence of Damage Formation in Metallic‐Intermetallic Mg–Al–Ca Composites at Elevated Temperature using Panoramic Image Analysis

A cast Mg‐4.65Al‐2.82Ca alloy with a microstructure containing an α‐Mg matrix is studied, which is reinforced with a C36 Laves phase skeleton. Such ternary alloys are targeted for elevated temperature applications in automotive engines, since they possess excellent creep properties. However, in application, the alloy may be subjected to a wide range of strain rates, and thus accelerated testing is often essential. It is, therefore, crucial to understand the effect of such rate variations. Herein, their impact on damage formation is focused. The analysis is based on high‐resolution panoramic imaging using scanning electron microscopy, combined with automated damage analysis using deep learning for object detection and classification (YOLOv5). It is found that with decreasing strain rate the dominant damage mechanism for a given strain level changes: at a strain rate of 5 × 10−4 s−1, the evolution of microcracks in the C36 Laves phase dominates damage formation. However, when the strain rate is decreased to 5 × 10−6 s−1, interface decohesion, becomes equally important. A change in crack orientation is also observed, indicating an increasing influence of plastic codeformation of the α‐Mg matrix and Laves phase. This transition is attributed to the leading damage mechanism to thermally activated processes at the interface.

An improved deep learning-based optimal object detection system from images

Computer vision technology for detecting objects in a complex environment often includes other key technologies, including pattern recognition, artificial intelligence, and digital image processing. It has been shown that Fast Convolutional Neural Networks (CNNs) with You Only Look Once (YOLO) is optimal for differentiating similar objects, constant motion, and low image quality. The proposed study aims to resolve these issues by implementing three different object detection algorithms—You Only Look Once (YOLO), Single Stage Detector (SSD), and Faster Region-Based Convolutional Neural Networks (R-CNN). This paper compares three different deep-learning object detection methods to find the best possible combination of feature and accuracy. The R-CNN object detection techniques are performed better than single-stage detectors like Yolo (You Only Look Once) and Single Shot Detector (SSD) in term of accuracy, recall, precision and loss.

Advancing Plastic Waste Classification and Recycling Efficiency: Integrating Image Sensors and Deep Learning Algorithms

Plastics, with their versatility and cost-effectiveness, have become indispensable materials across various industries. However, the improper disposal and mismanagement of plastic waste have led to significant environmental issues, including pollution, habitat destruction, and threats to wildlife. To address these challenges, numerous methods for plastic waste sorting and recycling have been developed. While conventional techniques like near-infrared spectroscopy (NIRS) have been effective to some extent, they face difficulties in accurately classifying chemically similar samples, such as polyethylene terephthalate (PET) and PET-glycol (PET-G), which have similar chemical compositions but distinct physical characteristics. This paper introduces an approach that adapts image sensors and deep learning object detection algorithms; specifically, the You Only Look Once (YOLO) model, to enhance plastic waste classification based on the shape of the waste. Unlike conventional methods that rely solely on spectral analysis, our methodology aims to significantly improve the accuracy and efficiency of classifying plastics, especially when dealing with materials having similar chemical compositions but differing physical attributes. The system developed using image sensors and the YOLO model proves to be not only effective but also scalable and adaptable for various industrial and environmental applications. In our experiments, the results are strikingly effective. We achieved a classification accuracy rate exceeding 91.7% mean Average Precision (mAP) in distinguishing between PET and PET-G, surpassing conventional techniques by a considerable margin. The implications of this research extend far and wide. By enhancing the accuracy of plastic waste sorting and reducing misclassification rates, we can significantly boost recycling efficiency. The proposed approach contributes to a more sustainable and efficient plastic waste management system, alleviating the strain on landfills and mitigating the environmental impact of plastic waste, contributing to a cleaner and more sustainable environment.

Background Instance-Based Copy-Paste Data Augmentation for Object Detection

In supervised deep learning object detection, the quantity of object information and annotation quality in a dataset affect model performance. To augment object detection datasets while maintaining contextual information between objects and backgrounds, we proposed a Background Instance-Based Copy-Paste (BIB-Copy-Paste) data augmentation model. We devised a method to generate background pseudo-labels for all object classes by calculating the similarity between object background features and image region features in Euclidean space. The background classifier, trained with these pseudo-labels, can guide copy-pasting to ensure contextual relevance. Several supervised object detectors were evaluated on the PASCAL VOC 2012 dataset, achieving a 1.1% average improvement in mean average precision. Ablation experiments with the BlitzNet object detector on the PASCAL VOC 2012 dataset showed an improvement of mAP by 1.19% using the proposed method, compared to a 0.18% improvement with random copy-paste. Images from the MS COCO dataset containing objects of the same classes as in PASCAL VOC 2012 were also selected for object pasting experiments. The contextual relevance of pasted objects demonstrated our model’s effectiveness and transferability between datasets with same class of objects.

Automated identification and counting of predated Ephestia kuehniella (Zeller) eggs using deep learning image analysis

Predation or kill rate of biological control agents is often used as a proxy to evaluate the efficacy of different species of natural enemies and under different conditions. For generalist predators and many egg parasitoids, eggs of the Mediterranean flour moth Ephestia kuehniella (Zeller) (Lepidoptera: Pyralidae) are used as factitious prey or hosts in laboratory and field experiments, as they are widely available from mass rearings of the biological control industry. Evaluating the predation or parasitism activity of natural enemies on E. kuehniella is a valuable tool for biological control practitioners around the world. However, manual assessments are laborious and prone to error, as observations may be subjective and depend on the individual performing the task. Here, we developed an automated protocol based on the deep learning object detection algorithm YOLOv5, for the accurate estimation of predated E. kuehniella eggs by piercing-sucking generalist predators. The application of the trained deep learning model achieved high precision (0.90) and recall (0.93) in the identification of predated eggs among intact eggs in our test experiment and was more accurate and faster than two independent observers that performed manual counting under a stereo microscope based on the mean absolute percentage error metric. Furthermore, a case study is presented where the predation activity of the generalist predators Orius laevigatus, Orius majusculus, Orius minutus, Nesidiocoris tenuis, Macrolophus pygmaeus, and Dicyphus errans is compared. Further applications and expansions of the protocol are discussed.

Extracting Masks from Herbarium Specimen Images Based on Object Detection and Image Segmentation Techniques

Extracting Masks from Herbarium Specimen Images Based on Object Detection and Image Segmentation Techniques

Can We Replace Real-World With Synthetic Data in Deep Learning-Based ADAS Algorithm Development?

Deep learning models require vast amounts of annotated training data. Gathering and annotating the data from the real world is an expensive and time-consuming process. Thus, synthetically generated data is being researched more and more. This paper tries to answer the question of whether and to what extent synthetically generated data can help in developing the advanced driver-assistance system (ADAS) algorithms for autonomous vehicles, for the object detection task based on deep learning. Two state-of-the-art deep learning object detectors were trained on various combinations of real-world and synthetic data. A total of 12 detectors were tested on real-world test images. Results show that synthetic data can contribute to better detector performance until a certain ratio between real-world and synthetic data is reached.

Machine vision-based automatic lamb identification and drinking activity in a commercial farm.

Using ear tags, farmers can track specific data for individual lambs such as age, medical records, body condition scores, genetic abnormalities; to make data-based decisions. However, automatic reading of ear tags using Radio Frequency Identification requires (a) an antenna, (b) a reader, (c) comparable reading standards; consequently, such a system can be expensive and impractical for a large group of lambs, especially in situations where animals are not required to have a compulsory Electronic identification, contrary to the case in Europe, where it is mandatory. Therefore, this paper proposes a machine vision system for indoor animals to identify individual lambs using existing ear tags. Using a camera that is installed such that the trough is visible, the drinking behaviour of the lambs can be automatically monitored. Data from different lamb groups in two different pens were collected. The identification algorithm includes a number of steps: (1) Detecting the lambs' face, and its ear tags in each image; (2) Cropping each ear tag image and discerning the digits on it to obtain the tag number; (3) Tracking each lamb throughout the visit using a tracking algorithm; (4) Recovering the ear tag number using an algorithm that incorporates a list of the ear tag numbers of the lambs in each pen, and the predictions for each lamb in each frame. The You Only Look Once deep learning object detection algorithm was applied to locate and localise the lamb's face and the digits in an image. The models' datasets contained 1160 and 2165 images for the training set, and 325 and 616 images for the validation set, respectively. The algorithm output includes the identity of each lamb that came to drink, and its duration. The identification system resulted in a total accuracy of 93% for the data tested, which consisted of approximately 900 visits to the drinking stations, and was collected in real time in a natural environment. The ground truth of each video of a visit was obtained by human observation by studying the video. We checked if there was indeed a visit to the water trough and if so we registered the ear tag number of each lamb whose head was above the water trough. Thus, identifying lambs in a commercial pen using a relatively inexpensive and easily installed system consisting of a RGB camera and a computer vision-based algorithm has potential for farm management.

Performance evaluation of deep learning models for the classification and identification of dental implants

Dental implant systems can be identified using image classification deep learning. However, investigations on the accuracy of classifying and identifying implant design through an object detection model are lacking. The purpose of this study was to evaluate the performance of an object detection deep learning model for classifying the implant designs of 103 types of implants. From panoramic radiographs, 14037 implant images were extracted. Implant designs were subdivided into 10 classes in the coronal, 13 in the middle, and 10 in the apical third. Classes with fewer than 50 images were excluded from the training dataset. Among the images, 80% were used as training data, and the remaining 20% as test data; the data were generated 3 times for 3-fold cross-validation (implant datasets 1, 2, and 3). Versions 5 and 7 of you only look once (YOLO) algorithm were used to train the model, and the mean average precision (mAP) was evaluated. Subsequently, data augmentation was performed using image processing and a real-enhanced super-resolution generative adversarial network, and the accuracy was re-evaluated using YOLOv7. The mAP of YOLOv7 in the 3 datasets was 0.931, 0.984, and 0.884, respectively, which were higher than the mAP of YOLOv5. After image processing in implant dataset-1, the mAP improved to 0.986 and, with the real-enhanced super-resolution generative adversarial network, to 0.988 and 0.986 at magnification ×2 and ×4, respectively. The object detection model for classifying implant designs found a high accuracy for 26 classes. The mAP of the model differed depending on the type of algorithm, image processing process, and detailed implant design.

Object Detection of Small Insects in Time-Lapse Camera Recordings.

As pollinators, insects play a crucial role in ecosystem management and world food production. However, insect populations are declining, necessitating efficient insect monitoring methods. Existing methods analyze video or time-lapse images of insects in nature, but analysis is challenging as insects are small objects in complex and dynamic natural vegetation scenes. In this work, we provide a dataset of primarily honeybees visiting three different plant species during two months of the summer. The dataset consists of 107,387 annotated time-lapse images from multiple cameras, including 9423 annotated insects. We present a method for detecting insects in time-lapse RGB images, which consists of a two-step process. Firstly, the time-lapse RGB images are preprocessed to enhance insects in the images. This motion-informed enhancement technique uses motion and colors to enhance insects in images. Secondly, the enhanced images are subsequently fed into a convolutional neural network (CNN) object detector. The method improves on the deep learning object detectors You Only Look Once (YOLO) and faster region-based CNN (Faster R-CNN). Using motion-informed enhancement, the YOLO detector improves the average micro F1-score from 0.49 to 0.71, and the Faster R-CNN detector improves the average micro F1-score from 0.32 to 0.56. Our dataset and proposed method provide a step forward for automating the time-lapse camera monitoring of flying insects.

Detecting the technology's evolutionary pathway using HiDS-trait-driven tech mining strategy

Detecting and identifying the future trends of disruptive technologies is crucial for both government R&D strategic planning and firms' practices. To monitor disruptive technology trends and grasp technological opportunities, this paper proposes an analytical framework that integrates appropriate tech mining and semantic analysis methods based on HiDS (high-dimensional and sparse)-trait patent abstract data and then visualizes the technology evolutionary pathway. Deep Learning technology is selected as a case study. In this case, tech mining methods are applied to analyze the Deep Learning technology evolutionary pathway and then refine Deep Learning technological prospects. Results indicate that Deep Learning has formulated three progressive clusters during its evolution, which strongly demonstrate the high application value of Deep Learning technology. In addition, Deep Learning technology will continue to maintain a rapid development stage, and Deep Learning Accelerator, Transfer Learning, Image Classification, Object Detection, and Deep Reinforcement Learning are the key technical topics in the future. Our study contributes to the technology forecast field by suggesting a data-trait-driven framework for HiDS data, which introduces data-trait-driven idea into technological evolution detection and combines it with technology life cycle theory, and helps experts make more precise and in-depth tech mining in technological evolution research.

PLATE: A perception-latency aware estimator

Target tracking is a popular problem with many potential applications. There has been a lot of effort on improving the quality of the detection of targets using cameras through different techniques. In general, with higher computational effort applied, i.e., a longer perception-latency, a better detection accuracy is obtained. However, it is not always useful to apply the longest perception-latency allowed, particularly when the environment does not require to and when the computational resources are shared between other tasks. In this work, we propose a new Perception-LATency aware Estimator (PLATE), which uses different perception configurations in different moments of time in order to optimize a certain performance measure. This measure takes into account a perception-latency and accuracy trade-off aiming for a good compromise between quality and resource usage. Compared to other heuristic frame-skipping techniques, PLATE comes with a formal complexity and optimality analysis. The advantages of PLATE are verified by several experiments including an evaluation over a standard benchmark with real data and using state of the art deep learning object detection methods for the perception stage.

MSDD-YOLOX: An enhanced YOLOX for real-time surface defect detection of oranges by type

Using an online high-throughput detection system for sorting oranges during the post-harvest process helped improve the commercialization level of the oranges industry. Surface defects on oranges created a poor first impression for consumers, making the rapid detection of oranges surface defects a primary concern for online sorting systems. However, due to variations in defect size and the visual similarity of different defects, there were still some challenges in detecting and identifying various surface defects on orange fruits based on their types. To address these challenges, this study first categorized surface defects on oranges into three major categories: deformity, scarring, and disease spot, based on their causes and potential post-harvest losses. Subsequently, to achieve real-time detection of orange surface defects on the orange sorting machine, a YOLOX-based real-time multi-type surface defect detection algorithm (MSDD-YOLOX) for oranges was proposed. This algorithm significantly improved the detection effectiveness of scarring at different scales by introducing neck network residual connections and cascading of the neck network. To address the issue of missed detections in texture-based defects and improve the regression of predicted bounding boxes, focal loss and Complete-IoU (CIoU) were employed in the algorithm. The results showed that MSDD-YOLOX achieved F1 values of 88.3 %, 80.4 %, and 92.7 % for the detection of deformity, scarring, and disease spot, respectively, with an overall detection F1 value of 90.8 %. These values represented improvements of 13.1 %, 10.2 %, 4.5 %, and 6.4 %, respectively, compared to the baseline model. Furthermore, compared to other deep learning object detectors, namely Faster RCNN, RetinaNet, FCOS, and Swin-Transformer, the proposed algorithm achieved optimal detection accuracy. Additionally, the MSDD-YOLOX model had a compact size of only 8.98 M, enabling real-time detection on the fruit grading line with an inference speed of up to 64.2FPS. Another innovation of this research was the external validation conducted on green oranges from Hainan and mandarins from Zhejiang. The results of external testing demonstrated that MSDD-YOLOX achieved overall F1 values of 90.6 % and 81.1 % for citrus fruits in these two regions, effectively proving the online deployment capability of MSDD-YOLOX and providing a robust solution for external defect detection in citrus fruits.

Analysis of Helmet Detection on Motor Drivers to Detect Traffic Violations Using the You Only Look Once Method (Yolov4)

According to statistical data, the number of deaths due to accidents in Indonesia in 2017 was 30,568 people. Efforts are being made to reduce traffic violations, especially helmet violations. Helmets that must be worn by Indonesian motorcyclists must comply with the Indonesian National Standard (SNI), but there are still many non-SNI helmets circulating. A possible solution for monitoring is the identification of motorbikes in traffic based on Deep Learning. In this study, the classification of helmets was carried out using the YO-LO (You Only Look Once) method. The SNI helmet detection system aims to make drivers more disciplined in completing their riding equipment, especially helmets with SNI because this system requires riders to wear helmets that comply with LLAJ or SNI (Indonesian National Standard) helmets before riding. Trending Machine Learning and Deep Learning conduct research to discover new methods and advanced architectures such as YOLO (You Only Look Once). YOLO is an object detection network architecture that is claimed to be the "fastest deep learning object detector" that prioritizes accuracy and speed. With YOLOv4, violations by motorbike riders can be detected in real-time and whether the riders recorded on the camera are directly wearing SNI helmets, non-SNI helmets or not wearing helmets. The best accuracy for real-time motorcyclist violations with YOLOv4 is the best mAP value of 99.69%

CAM-YOLO: tomato detection and classification based on improved YOLOv5 using combining attention mechanism.

One of the key elements in maintaining the consistent marketing of tomato fruit is tomato quality. Since ripeness is the most important factor for tomato quality in the viewpoint of consumers, determining the stages of tomato ripeness is a fundamental industrial concern with regard to tomato production to obtain a high quality product. Since tomatoes are one of the most important crops in the world, automatic ripeness evaluation of tomatoes is a significant study topic as it may prove beneficial in ensuring an optimal production of high-quality product, increasing profitability. This article explores and categorises the various maturity/ripeness phases to propose an automated multi-class classification approach for tomato ripeness testing and evaluation. Object detection is the critical component in a wide variety of computer vision problems and applications such as manufacturing, agriculture, medicine, and autonomous driving. Due to the tomato fruits' complex identification background, texture disruption, and partial occlusion, the classic deep learning object detection approach (YOLO) has a poor rate of success in detecting tomato fruits. To figure out these issues, this article proposes an improved YOLOv5 tomato detection algorithm. The proposed algorithm CAM-YOLO uses YOLOv5 for feature extraction, target identification and Convolutional Block Attention Module (CBAM). The CBAM is added to the CAM-YOLO to focus the model on improving accuracy. Finally, non-maximum suppression and distance intersection over union (DIoU) are applied to enhance the identification of overlapping objects in the image. Several images from the dataset were chosen for testing to assess the model's performance, and the detection performance of the CAM-YOLO and standard YOLOv5 models under various conditions was compared. The experimental results affirms that CAM-YOLO algorithm is efficient in detecting the overlapped and small tomatoes with an average precision of 88.1%.

Deep learning system assisted detection and localization of lumbar spondylolisthesis.

Objective: Explore a new deep learning (DL) object detection algorithm for clinical auxiliary diagnosis of lumbar spondylolisthesis and compare it with doctors' evaluation to verify the effectiveness and feasibility of the DL algorithm in the diagnosis of lumbar spondylolisthesis. Methods: Lumbar lateral radiographs of 1,596 patients with lumbar spondylolisthesis from three medical institutions were collected, and senior orthopedic surgeons and radiologists jointly diagnosed and marked them to establish a database. These radiographs were randomly divided into a training set (n = 1,117), a validation set (n = 240), and a test set (n = 239) in a ratio of 0.7 : 0.15: 0.15. We trained two DL models for automatic detection of spondylolisthesis and evaluated their diagnostic performance by PR curves, areas under the curve, precision, recall, F1-score. Then we chose the model with better performance and compared its results with professionals' evaluation. Results: A total of 1,780 annotations were marked for training (1,242), validation (263), and test (275). The Faster Region-based Convolutional Neural Network (R-CNN) showed better precision (0.935), recall (0.935), and F1-score (0.935) in the detection of spondylolisthesis, which outperformed the doctor group with precision (0.927), recall (0.892), f1-score (0.910). In addition, with the assistance of the DL model, the precision of the doctor group increased by 4.8%, the recall by 8.2%, the F1-score by 6.4%, and the average diagnosis time per plain X-ray was shortened by 7.139s. Conclusion: The DL detection algorithm is an effective method for clinical diagnosis of lumbar spondylolisthesis. It can be used as an assistant expert to improve the accuracy of lumbar spondylolisthesis diagnosis and reduce the clinical workloads.

Artificial Intelligence Techniques for Automatic Detection of Peri-implant Marginal Bone Remodeling in Intraoral Radiographs

Peri-implantitis can cause marginal bone remodeling around implants. The aim is to develop an automatic image processing approach based on two artificial intelligence (AI) techniques in intraoral (periapical and bitewing) radiographs to assist dentists in determining bone loss. The first is a deep learning (DL) object-detector (YOLOv3) to roughly identify (no exact localization is required) two objects: prosthesis (crown) and implant (screw). The second is an image understanding-based (IU) process to fine-tune lines on screw edges and to identify significant points (intensity bone changes, intersections between screw and crown). Distances between these points are used to compute bone loss. A total of 2920 radiographs were used for training (50%) and testing (50%) the DL process. The mAP@0.5 metric is used for performance evaluation of DL considering periapical/bitewing and screws/crowns in upper and lower jaws, with scores ranging from 0.537 to 0.898 (sufficient because DL only needs an approximation). The IU performance is assessed with 50% of the testing radiographs through the t test statistical method, obtaining p values of 0.0106 (line fitting) and 0.0213 (significant point detection). The IU performance is satisfactory, as these values are in accordance with the statistical average/standard deviation in pixels for line fitting (2.75/1.01) and for significant point detection (2.63/1.28) according to the expert criteria of dentists, who establish the ground-truth lines and significant points. In conclusion, AI methods have good prospects for automatic bone loss detection in intraoral radiographs to assist dental specialists in diagnosing peri-implantitis.

Soft computing in business: exploring current research and outlining future research directions

PurposeThe primary aim of this study is to detail the use of soft computing techniques in business and management research. Its objectives are as follows: to conduct a comprehensive scientometric analysis of publications in the field of soft computing, to explore the evolution of keywords, to identify key research themes and latent topics and to map the intellectual structure of soft computing in the business literature.Design/methodology/approachThis research offers a comprehensive overview of the field by synthesising 43 years (1980–2022) of soft computing research from the Scopus database. It employs descriptive analysis, topic modelling (TM) and scientometric analysis.FindingsThis study's co-citation analysis identifies three primary categories of research in the field: the components, the techniques and the benefits of soft computing. Additionally, this study identifies 16 key study themes in the soft computing literature using TM, including decision-making under uncertainty, multi-criteria decision-making (MCDM), the application of deep learning in object detection and fault diagnosis, circular economy and sustainable development and a few others.Practical implicationsThis analysis offers a valuable understanding of soft computing for researchers and industry experts and highlights potential areas for future research.Originality/valueThis study uses scientific mapping and performance indicators to analyse a large corpus of 4,512 articles in the field of soft computing. It makes significant contributions to the intellectual and conceptual framework of soft computing research by providing a comprehensive overview of the literature on soft computing literature covering a period of four decades and identifying significant trends and topics to direct future research.

Few-Shot Object Detection in Remote Sensing Imagery via Fuse Context Dependencies and Global Features

The rapid development of Earth observation technology has promoted the continuous accumulation of images in the field of remote sensing. However, a large number of remote sensing images still lack manual annotations of objects, which makes the strongly supervised deep learning object detection method not widely used, as it lacks generalization ability for unseen object categories. Considering the above problems, this study proposes a few-shot remote sensing image object detection method that integrates context dependencies and global features. The method can be used to fine-tune the model with a small number of sample annotations based on the model trained in the base class, as a way to enhance the detection capability of new object classes. The method proposed in this study consists of three main modules, namely, the meta-feature extractor (ME), reweighting module (RM), and feature fusion module (FFM). These three modules are respectively used to enhance the context dependencies of the query set features, improve the global features of the support set that contains annotations, and finally fuse the query set features and support set features. The baseline of the meta-feature extractor of the entire framework is based on the optimized YOLOv5 framework. The reweighting module of the support set feature extraction is based on a simple convolutional neural network (CNN) framework, and the foreground feature enhancement of the support sets was made in the preprocessing stage. This study achieved beneficial results in the two benchmark datasets NWPU VHR-10 and DIOR. Compared with the comparison methods, the proposed method achieved the best performance in the object detection of the base class and the novel class.

Image-Processing-Based Subway Tunnel Crack Detection System.

With the increase in urban rail transit construction, instances of tunnel disease are on the rise, and cracks have become the focus of tunnel maintenance and management. Therefore, it is essential to carry out crack detection in a timely and efficient manner to not only prolong the service life of the tunnel but also reduce the incidence of accidents. In this paper, the design and structure of a tunnel crack detection system are analyzed. On this basis, this paper proposes a new method for crack identification and feature detection using image processing technology. This method fully considers the characteristics of tunnel images and the combination of these characteristics with deep learning, while a deep convolutional network (Single-Shot MultiBox Detector (SSD)) is proposed based on deep learning for object detection in complex images. The experimental results show that the test set accuracy and training set accuracy of the support vector machine (SVM) in the classification comparison test are up to 88% and 87.8%, respectively; while the test accuracy of Alexnet's deep convolutional neural network-based classification and identification is up to 96.7%, and the training set accuracy is up to 97.5%. It can be seen that this deep convolutional network recognition algorithm based on deep learning and image processing is better and more suitable for the detection of cracks in subway tunnels.