Hong-Dar Lin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Hong-Dar Lin has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
- Chou-Hsien Lin
Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin, Austin, TX 78712-0273, USA
11 shared publications Yuan-Shyi Peter Chiu
Faculty of Industrial Engineering & Management, Chaoyang University of Technology, Taichung City 413, Taiwan
8 shared publicationsSinga Wang Chiu
Chaoyang Univ Technol, Dept Business Adm, Taichung, Taiwan
3 shared publicationsChih-Hao Chien
Chaoyang Univ. of Technol., Taichung
2 shared publicationsChou-Hsien Lin
Department of Civil, Architectural, and Environmental Engineering, The University of Texas at Austin, 301, East Dean Keeton Street, Austin, TX, 78712-0273, USA
2 shared publicationsDuan-Cheng Ho
Department of Industrial Engineering and Management, Chaoyang University of Technology, Taichung County, Taiwan
2 shared publications
Latest Publications
- Open Access
Vision-Based Automated Inspection of Box Meals for Food Portion Defects and Foreign Object Detection
Automated inspection of prepared meals is important for improving food safety, quality assurance, and production efficiency. However, vision-based inspection remains challenging because boxed meals contain multiple adjacent food items with irregular shapes, varying portion sizes, and visually similar appearances, while oily surfaces may introduce specular reflections that degrade image quality. This study presents a vision-based framework for multi-object recognition and quantitative defect analysis in complex meal images, with Chinese-style lunch boxes used as representative test samples. The framework integrates region-of-interest (ROI) extraction, fixed-grid regional feature representation, deep neural network (DNN) classification, flood-filling post-processing, and empirical food-quantity thresholds. The lunch-box ROI is first extracted using the Hough transform, followed by median filtering to suppress reflection noise. The ROI is partitioned into 6 × 6 regions, from which the mean and standard deviation of RGB, HSV, and CIE Lab* color components are extracted and classified using a DNN. Flood filling is subsequently applied to refine the classification results, and category-specific empirical thresholds are used to identify missing food items and insufficient portions. Foreign objects are detected as an additional abnormal category. Under the evaluated experimental conditions, the proposed framework achieved an overall image classification rate (CR) of 96.45%, a defective-image detection rate (1−β) of 97.76%, a normal-image false alarm rate (α) of 5.34%, and a defective-image misclassification rate (γ) of 0.82%. Sensitivity experiments involving illumination variation, two lunch-box configurations with different food compositions, and conveyor-based image acquisition further demonstrated the feasibility and stability of the framework under the evaluated laboratory and prototype conditions. These findings support the feasibility of the proposed lightweight framework for vision-based quality inspection of representative boxed meals, while broader validation across meal types, contaminants, and industrial production environments remains necessary.
Open AccessVision-Based Automated Inspection of Box Meals for Food Portion Defects and Foreign Object Detection
Automated inspection of prepared meals is important for improving food safety, quality assurance, and production efficiency. However, vision-based inspection remains challenging because boxed meals contain multiple adjacent food items with irregular shapes, varying portion sizes, and visually similar appearances, while oily surfaces may introduce specular reflections that degrade image quality. This study presents a vision-based framework for multi-object recognition and quantitative defect analysis in complex meal images, with Chinese-style lunch boxes used as representative test samples. The framework integrates region-of-interest (ROI) extraction, fixed-grid regional feature representation, deep neural network (DNN) classification, flood-filling post-processing, and empirical food-quantity thresholds. The lunch-box ROI is first extracted using the Hough transform, followed by median filtering to suppress reflection noise. The ROI is partitioned into 6 × 6 regions, from which the mean and standard deviation of RGB, HSV, and CIE Lab* color components are extracted and classified using a DNN. Flood filling is subsequently applied to refine the classification results, and category-specific empirical thresholds are used to identify missing food items and insufficient portions. Foreign objects are detected as an additional abnormal category. Under the evaluated experimental conditions, the proposed framework achieved an overall image classification rate (CR) of 96.45%, a defective-image detection rate (1−β) of 97.76%, a normal-image false alarm rate (α) of 5.34%, and a defective-image misclassification rate (γ) of 0.82%. Sensitivity experiments involving illumination variation, two lunch-box configurations with different food compositions, and conveyor-based image acquisition further demonstrated the feasibility and stability of the framework under the evaluated laboratory and prototype conditions. These findings support the feasibility of the proposed lightweight framework for vision-based quality inspection of representative boxed meals, while broader validation across meal types, contaminants, and industrial production environments remains necessary.
- Open Access
Vision-Based Quality Grading of Beef Steaks Using Marbling Distribution Analysis and Lean Meat Color Classification
This study proposes a vision-based framework for automated inspection and quality grading of beef steaks by integrating fat marbling distribution analysis and lean-meat color evaluation. In frozen beef products, surface frost often generates specular reflections that resemble both fat and lean regions, thereby reducing segmentation accuracy. To address this challenge, a sequential and interpretable analytical framework is developed. First, homomorphic filtering is applied to suppress frost-induced illumination artifacts, followed by curvelet transform combined with square-ring filtering to separate fat and lean regions based on their multi-scale and directional characteristics. For marbling analysis, the convex hull, skeleton, and principal axis of the steak are extracted, and a chi-square goodness-of-fit test is performed within eight predefined regions to quantitatively evaluate marbling distribution uniformity and identify localized fat accumulation. For lean-meat evaluation, RGB color features are extracted and classified using a Support Vector Machine (SVM) to determine redness levels. The resulting marbling and color information are subsequently integrated through a weighted grading strategy to estimate the final quality grade. Experimental results demonstrate a fat detection rate of 92.68%, a false-positive rate of 4.97%, and a correct classification rate of 94.09% for fat segmentation, while the SVM-based lean-meat color classifier achieves an accuracy of 96.67%. Furthermore, the proposed grading framework attains an overall grading accuracy of 90.38%, showing strong agreement with human evaluation.
Open AccessVision-Based Quality Grading of Beef Steaks Using Marbling Distribution Analysis and Lean Meat Color Classification
This study proposes a vision-based framework for automated inspection and quality grading of beef steaks by integrating fat marbling distribution analysis and lean-meat color evaluation. In frozen beef products, surface frost often generates specular reflections that resemble both fat and lean regions, thereby reducing segmentation accuracy. To address this challenge, a sequential and interpretable analytical framework is developed. First, homomorphic filtering is applied to suppress frost-induced illumination artifacts, followed by curvelet transform combined with square-ring filtering to separate fat and lean regions based on their multi-scale and directional characteristics. For marbling analysis, the convex hull, skeleton, and principal axis of the steak are extracted, and a chi-square goodness-of-fit test is performed within eight predefined regions to quantitatively evaluate marbling distribution uniformity and identify localized fat accumulation. For lean-meat evaluation, RGB color features are extracted and classified using a Support Vector Machine (SVM) to determine redness levels. The resulting marbling and color information are subsequently integrated through a weighted grading strategy to estimate the final quality grade. Experimental results demonstrate a fat detection rate of 92.68%, a false-positive rate of 4.97%, and a correct classification rate of 94.09% for fat segmentation, while the SVM-based lean-meat color classifier achieves an accuracy of 96.67%. Furthermore, the proposed grading framework attains an overall grading accuracy of 90.38%, showing strong agreement with human evaluation.
- Open Access
An Automated Vision-Based Inspection System for Metallic Lock Surface Defects Using a Transformer-Enhanced U-Net
Surface defect inspection of metallic lock components remains challenging due to strong specular reflections, low-contrast defect patterns, and geometric variability, which limit the consistency of manual inspection and conventional automated optical inspection (AOI) systems. This study presents an integrated visual inspection framework that combines controlled image acquisition with deep learning-based semantic segmentation to enable reliable and repeatable defect detection. A standardized rotational fixture with ring illumination was developed to stabilize imaging geometry, reduce reflection variability, and support consistent multi-view acquisition. A region-of-interest (ROI) masking strategy was further applied to suppress background interference and isolate the effective inspection region. At the algorithmic level, a Transformer-enhanced U-Net (TransU-Net) architecture was employed to jointly model local spatial features and global contextual dependencies, thereby improving boundary delineation and the detection of irregular surface anomalies. In addition, a boundary-aware weighted evaluation scheme was introduced to provide a more robust and application-relevant assessment by accounting for annotation uncertainty near defect edges. Experimental results demonstrate that the proposed method achieved an F1-score of 85.15%, with an average inference time of 0.3357 s per image for model prediction. Considering additional processes such as multi-view image acquisition, mechanical rotation, and preprocessing, the overall system-level inspection time is expected to be on the order of seconds per component in practical deployment.
Open AccessAn Automated Vision-Based Inspection System for Metallic Lock Surface Defects Using a Transformer-Enhanced U-Net
Surface defect inspection of metallic lock components remains challenging due to strong specular reflections, low-contrast defect patterns, and geometric variability, which limit the consistency of manual inspection and conventional automated optical inspection (AOI) systems. This study presents an integrated visual inspection framework that combines controlled image acquisition with deep learning-based semantic segmentation to enable reliable and repeatable defect detection. A standardized rotational fixture with ring illumination was developed to stabilize imaging geometry, reduce reflection variability, and support consistent multi-view acquisition. A region-of-interest (ROI) masking strategy was further applied to suppress background interference and isolate the effective inspection region. At the algorithmic level, a Transformer-enhanced U-Net (TransU-Net) architecture was employed to jointly model local spatial features and global contextual dependencies, thereby improving boundary delineation and the detection of irregular surface anomalies. In addition, a boundary-aware weighted evaluation scheme was introduced to provide a more robust and application-relevant assessment by accounting for annotation uncertainty near defect edges. Experimental results demonstrate that the proposed method achieved an F1-score of 85.15%, with an average inference time of 0.3357 s per image for model prediction. Considering additional processes such as multi-view image acquisition, mechanical rotation, and preprocessing, the overall system-level inspection time is expected to be on the order of seconds per component in practical deployment.
- Open Access
Smart Image-Based Deep Learning System for Automated Quality Grading of Phalaenopsis Seedlings in Outsourced Production
Phalaenopsis orchids are one of Taiwan’s key floral export products, and maintaining consistent quality is crucial for international competitiveness. To improve production efficiency, many orchid farms outsource the early flask seedling stage to contract growers, who raise the plants to the 2.5-inch potted seedling stage before returning them for further greenhouse cultivation. Traditionally, the quality of these outsourced seedlings is evaluated manually by inspectors who visually detect defects and assign quality grades based on experience, a process that is time-consuming and subjective. This study introduces a smart image-based deep learning system for automatic quality grading of Phalaenopsis potted seedlings, combining computer vision, deep learning, and machine learning techniques to replace manual inspection. The system uses YOLOv8 and YOLOv10 models for defect and root detection, along with SVM and Random Forest classifiers for defect counting and grading. It employs a dual-view imaging approach, utilizing top-view RGB-D images to capture spatial leaf structures and multi-angle side-view RGB images to assess leaf and root conditions. Two grading strategies are developed: a three-stage hierarchical method that offers interpretable diagnostic results and a direct grading method for fast, end-to-end quality prediction. Performance comparisons and ablation studies show that using RGB-D top-view images and optimal viewing-angle combinations significantly improve grading accuracy. The system achieves F1-scores of 84.44% (three-stage) and 90.44% (direct), demonstrating high reliability and strong potential for automated quality assessment and export inspection in the orchid industry.
Open AccessSmart Image-Based Deep Learning System for Automated Quality Grading of Phalaenopsis Seedlings in Outsourced Production
Phalaenopsis orchids are one of Taiwan’s key floral export products, and maintaining consistent quality is crucial for international competitiveness. To improve production efficiency, many orchid farms outsource the early flask seedling stage to contract growers, who raise the plants to the 2.5-inch potted seedling stage before returning them for further greenhouse cultivation. Traditionally, the quality of these outsourced seedlings is evaluated manually by inspectors who visually detect defects and assign quality grades based on experience, a process that is time-consuming and subjective. This study introduces a smart image-based deep learning system for automatic quality grading of Phalaenopsis potted seedlings, combining computer vision, deep learning, and machine learning techniques to replace manual inspection. The system uses YOLOv8 and YOLOv10 models for defect and root detection, along with SVM and Random Forest classifiers for defect counting and grading. It employs a dual-view imaging approach, utilizing top-view RGB-D images to capture spatial leaf structures and multi-angle side-view RGB images to assess leaf and root conditions. Two grading strategies are developed: a three-stage hierarchical method that offers interpretable diagnostic results and a direct grading method for fast, end-to-end quality prediction. Performance comparisons and ablation studies show that using RGB-D top-view images and optimal viewing-angle combinations significantly improve grading accuracy. The system achieves F1-scores of 84.44% (three-stage) and 90.44% (direct), demonstrating high reliability and strong potential for automated quality assessment and export inspection in the orchid industry.
- Open Access
A Deep Learning-Based Sensing System for Identifying Salmon and Rainbow Trout Meat and Grading Freshness for Consumer Protection
Seafood fraud, such as mislabeling low-cost rainbow trout as premium salmon, poses serious food safety risks and damages consumer rights. To address this growing concern, this study develops a deep learning-based, smartphone-compatible sensing system for fish meat identification and salmon freshness grading. By providing consumers with real-time, image-based verification tools, the system supports informed purchasing decisions and enhances food safety. The system adopts a two-stage design: first classifying fish meat types, then grading salmon freshness into three levels based on visual cues. An improved DenseNet121 architecture, enhanced with global average pooling, dropout layers, and a customized output layer, improves accuracy and reduces overfitting, while transfer learning with partial layer freezing enhances efficiency by reducing training time without significant accuracy loss. Experimental results show that the two-stage method outperforms the one-stage approach and several baseline models, achieving robust accuracy in both classification and grading tasks. Sensitivity analysis demonstrates resilience to blur and camera tilt, though real-world adaptability under diverse lighting and packaging conditions remains a challenge. Overall, the proposed system represents a practical, consumer-oriented tool for seafood authentication and freshness evaluation, with potential to enhance food safety and consumer protection.
Open AccessA Deep Learning-Based Sensing System for Identifying Salmon and Rainbow Trout Meat and Grading Freshness for Consumer Protection
Seafood fraud, such as mislabeling low-cost rainbow trout as premium salmon, poses serious food safety risks and damages consumer rights. To address this growing concern, this study develops a deep learning-based, smartphone-compatible sensing system for fish meat identification and salmon freshness grading. By providing consumers with real-time, image-based verification tools, the system supports informed purchasing decisions and enhances food safety. The system adopts a two-stage design: first classifying fish meat types, then grading salmon freshness into three levels based on visual cues. An improved DenseNet121 architecture, enhanced with global average pooling, dropout layers, and a customized output layer, improves accuracy and reduces overfitting, while transfer learning with partial layer freezing enhances efficiency by reducing training time without significant accuracy loss. Experimental results show that the two-stage method outperforms the one-stage approach and several baseline models, achieving robust accuracy in both classification and grading tasks. Sensitivity analysis demonstrates resilience to blur and camera tilt, though real-world adaptability under diverse lighting and packaging conditions remains a challenge. Overall, the proposed system represents a practical, consumer-oriented tool for seafood authentication and freshness evaluation, with potential to enhance food safety and consumer protection.
- Open Access
Smartphone-Based Sensing System for Identifying Artificially Marbled Beef Using Texture and Color Analysis to Enhance Food Safety
Beef fat injection technology, used to enhance the perceived quality of lower-grade meat, often results in artificially marbled beef that mimics the visual traits of Wagyu, characterized by dense fat distribution. This practice, driven by the high cost of Wagyu and the affordability of fat-injected beef, has led to the proliferation of mislabeled “Wagyu-grade” products sold at premium prices, posing potential food safety risks such as allergen exposure or consumption of unverified additives, which can adversely affect consumer health. Addressing this, this study introduces a smart sensing system integrated with handheld mobile devices, enabling consumers to capture beef images during purchase for real-time health-focused assessment. The system analyzes surface texture and color, transmitting data to a server for classification to determine if the beef is artificially marbled, thus supporting informed dietary choices and reducing health risks. Images are processed by applying a region of interest (ROI) mask to remove background noise, followed by partitioning into grid blocks. Local binary pattern (LBP) texture features and RGB color features are extracted from these blocks to characterize surface properties of three beef types (Wagyu, regular, and fat-injected). A support vector machine (SVM) model classifies the blocks, with the final image classification determined via majority voting. Experimental results reveal that the system achieves a recall rate of 95.00% for fat-injected beef, a misjudgment rate of 1.67% for non-fat-injected beef, a correct classification rate (CR) of 93.89%, and an F1-score of 95.80%, demonstrating its potential as a human-centered healthcare tool for ensuring food safety and transparency.
Open AccessSmartphone-Based Sensing System for Identifying Artificially Marbled Beef Using Texture and Color Analysis to Enhance Food Safety
Beef fat injection technology, used to enhance the perceived quality of lower-grade meat, often results in artificially marbled beef that mimics the visual traits of Wagyu, characterized by dense fat distribution. This practice, driven by the high cost of Wagyu and the affordability of fat-injected beef, has led to the proliferation of mislabeled “Wagyu-grade” products sold at premium prices, posing potential food safety risks such as allergen exposure or consumption of unverified additives, which can adversely affect consumer health. Addressing this, this study introduces a smart sensing system integrated with handheld mobile devices, enabling consumers to capture beef images during purchase for real-time health-focused assessment. The system analyzes surface texture and color, transmitting data to a server for classification to determine if the beef is artificially marbled, thus supporting informed dietary choices and reducing health risks. Images are processed by applying a region of interest (ROI) mask to remove background noise, followed by partitioning into grid blocks. Local binary pattern (LBP) texture features and RGB color features are extracted from these blocks to characterize surface properties of three beef types (Wagyu, regular, and fat-injected). A support vector machine (SVM) model classifies the blocks, with the final image classification determined via majority voting. Experimental results reveal that the system achieves a recall rate of 95.00% for fat-injected beef, a misjudgment rate of 1.67% for non-fat-injected beef, a correct classification rate (CR) of 93.89%, and an F1-score of 95.80%, demonstrating its potential as a human-centered healthcare tool for ensuring food safety and transparency.
- Open Access
Smartphone-Based Deep Learning System for Detecting Ractopamine-Fed Pork Using Visual Classification Techniques
Ractopamine, a beta-agonist used to enhance lean meat yield, poses health risks with excessive consumption. To comply with global trade policies, Taiwan permits imports of North American (USA. and Canadian) pork containing ractopamine, raising concerns over unclear labeling and potential misidentification as Taiwan pork. Given the high demand for pork, consumers need a reliable way to verify meat authenticity. To address this issue, this study proposes a smartphone-based visual detection system for meat cut and pork origin classification, extending to ractopamine detection. Consumers can use mobile devices in retail settings to analyze pork images and make informed decisions. The system employs a three-stage process: first, applying a black elliptical mask to extract the outer ROI (region of interest) for meat cut classification; then, using a black square mask to obtain the inner ROI for pork origin classification; and finally, determining ractopamine presence in North American pork. Experimental results demonstrate MobileNet’s superior accuracy and efficiency, achieving a 96% CR (classification rate) for meat cut classification, a 79.11% average CR and 90.25% F1 score for pork origin classification, and an 80.67% average CR and 80.56% F1 score for ractopamine detection. These findings confirm the system’s effectiveness in enhancing meat authenticity verification and market transparency.
Open AccessSmartphone-Based Deep Learning System for Detecting Ractopamine-Fed Pork Using Visual Classification Techniques
Ractopamine, a beta-agonist used to enhance lean meat yield, poses health risks with excessive consumption. To comply with global trade policies, Taiwan permits imports of North American (USA. and Canadian) pork containing ractopamine, raising concerns over unclear labeling and potential misidentification as Taiwan pork. Given the high demand for pork, consumers need a reliable way to verify meat authenticity. To address this issue, this study proposes a smartphone-based visual detection system for meat cut and pork origin classification, extending to ractopamine detection. Consumers can use mobile devices in retail settings to analyze pork images and make informed decisions. The system employs a three-stage process: first, applying a black elliptical mask to extract the outer ROI (region of interest) for meat cut classification; then, using a black square mask to obtain the inner ROI for pork origin classification; and finally, determining ractopamine presence in North American pork. Experimental results demonstrate MobileNet’s superior accuracy and efficiency, achieving a 96% CR (classification rate) for meat cut classification, a 79.11% average CR and 90.25% F1 score for pork origin classification, and an 80.67% average CR and 80.56% F1 score for ractopamine detection. These findings confirm the system’s effectiveness in enhancing meat authenticity verification and market transparency.
- Open Access
A Deep Transfer Learning-Based Visual Inspection System for Assembly Defects in Similar Types of Manual Tool Products
This study introduces an advanced inspection system for manual tool assembly, focusing on defect detection and classification in flex-head ratchet wrenches as a modern alternative to traditional inspection methods. Using a deep learning R-CNN approach with transfer learning, specifically utilizing the AlexNet architecture, the system accurately identifies and classifies assembly defects across similar tools. This study demonstrates how a pre-trained defect detection model for older manual tool models can be efficiently adapted to new models with only moderate amounts of new samples and fine-tuning. Experimental evaluations at three assembly stations show that the AlexNet model achieves a classification accuracy of 98.67% at the station with the highest defect variety, outperforming the R-CNN model with randomly initialized weights. Even with a 40% reduction in sample size for new products, the AlexNet model maintains a classification accuracy of 98.66%. Additionally, compared to R-CNN, it improves average effectiveness by 9% and efficiency by 26% across all stations. A sensitivity analysis further reveals that the proposed method reduces training samples by 50% at 50% similarity while enhancing effectiveness by 13.06% and efficiency by 5.31%.
Open AccessA Deep Transfer Learning-Based Visual Inspection System for Assembly Defects in Similar Types of Manual Tool Products
This study introduces an advanced inspection system for manual tool assembly, focusing on defect detection and classification in flex-head ratchet wrenches as a modern alternative to traditional inspection methods. Using a deep learning R-CNN approach with transfer learning, specifically utilizing the AlexNet architecture, the system accurately identifies and classifies assembly defects across similar tools. This study demonstrates how a pre-trained defect detection model for older manual tool models can be efficiently adapted to new models with only moderate amounts of new samples and fine-tuning. Experimental evaluations at three assembly stations show that the AlexNet model achieves a classification accuracy of 98.67% at the station with the highest defect variety, outperforming the R-CNN model with randomly initialized weights. Even with a 40% reduction in sample size for new products, the AlexNet model maintains a classification accuracy of 98.66%. Additionally, compared to R-CNN, it improves average effectiveness by 9% and efficiency by 26% across all stations. A sensitivity analysis further reveals that the proposed method reduces training samples by 50% at 50% similarity while enhancing effectiveness by 13.06% and efficiency by 5.31%.
- Open Access
Visual defect inspection of touch screens using multi-angle filtering in curvelet domain
- Open Access
Utilizing Deep Learning for Defect Inspection in Hand Tool Assembly
The integrity of product assembly in the precision assembly industry significantly influences the quality of the final products. During the assembly process, products may acquire assembly defects due to personnel oversight. A severe assembly defect could impair the product’s normal function and potentially cause loss of life or property for the user. For workpiece defect inspection, there is limited discussion on the simultaneous detection of the primary kinds of assembly anomaly (missing parts, misplaced parts, foreign objects, and extra parts). However, these assembly anomalies account for most customer complaints in the traditional hand tool industry. This is because no equipment can comprehensively inspect major assembly defects, and inspections rely solely on professionals using simple tools and their own experience. Thus, this study proposes an automated visual inspection system to achieve defect inspection in hand tool assembly. This study samples the work-in-process from three assembly stations in the ratchet wrench assembly process; an investigation of 28 common assembly defect types is presented, covering the 4 kinds of assembly anomaly in the assembly operation; also, this study captures sample images of various assembly defects for the experiments. First, the captured images are filtered to eliminate surface reflection noise from the workpiece; then, a circular mask is given at the assembly position to extract the ROI area; next, the filtered ROI images are used to create a defect-type label set using manual annotation; after this, the R-CNN series network models are applied to object feature extraction and classification; finally, they are compared with other object detection models to identify which inspection model has the better performance. The experimental results show that, if each station uses the best model for defect inspection, it can effectively detect and classify defects. The average defect detection rate (1-β) of each station is 92.64%, the average misjudgment rate (α) is 6.68%, and the average correct classification rate (CR) is 88.03%.
Open AccessUtilizing Deep Learning for Defect Inspection in Hand Tool Assembly
The integrity of product assembly in the precision assembly industry significantly influences the quality of the final products. During the assembly process, products may acquire assembly defects due to personnel oversight. A severe assembly defect could impair the product’s normal function and potentially cause loss of life or property for the user. For workpiece defect inspection, there is limited discussion on the simultaneous detection of the primary kinds of assembly anomaly (missing parts, misplaced parts, foreign objects, and extra parts). However, these assembly anomalies account for most customer complaints in the traditional hand tool industry. This is because no equipment can comprehensively inspect major assembly defects, and inspections rely solely on professionals using simple tools and their own experience. Thus, this study proposes an automated visual inspection system to achieve defect inspection in hand tool assembly. This study samples the work-in-process from three assembly stations in the ratchet wrench assembly process; an investigation of 28 common assembly defect types is presented, covering the 4 kinds of assembly anomaly in the assembly operation; also, this study captures sample images of various assembly defects for the experiments. First, the captured images are filtered to eliminate surface reflection noise from the workpiece; then, a circular mask is given at the assembly position to extract the ROI area; next, the filtered ROI images are used to create a defect-type label set using manual annotation; after this, the R-CNN series network models are applied to object feature extraction and classification; finally, they are compared with other object detection models to identify which inspection model has the better performance. The experimental results show that, if each station uses the best model for defect inspection, it can effectively detect and classify defects. The average defect detection rate (1-β) of each station is 92.64%, the average misjudgment rate (α) is 6.68%, and the average correct classification rate (CR) is 88.03%.