• Korean Journal of Remote Sensing
• /
• v.40 no.1
• /
• pp.71-80
• /
• 2024

In this study, a proposed ensemble learning technique aims to enhance the semantic segmentation performance of images captured by Unmanned Aerial Vehicles (UAVs). With the increasing use of UAVs in fields such as urban planning, there has been active development of techniques utilizing deep learning segmentation methods for land cover segmentation. The study suggests a method that utilizes prominent segmentation models, namely U-Net, DeepLabV3, and Fully Convolutional Network (FCN), to improve segmentation prediction performance. The proposed approach integrates training loss, validation accuracy, and class score of the three segmentation models to enhance overall prediction performance. The method was applied and evaluated on a land cover segmentation problem involving seven classes: buildings,roads, parking lots, fields, trees, empty spaces, and areas with unspecified labels, using images captured by UAVs. The performance of the ensemble model was evaluated by mean Intersection over Union (mIoU), and the results of comparing the proposed ensemble model with the three existing segmentation methods showed that mIoU performance was improved. Consequently, the study confirms that the proposed technique can enhance the performance of semantic segmentation models.

• Smart Structures and Systems
• /
• v.31 no.4
• /
• pp.335-349
• /
• 2023

Bridges constantly undergo deterioration and damage, the most common ones being concrete damage and exposed rebar. Periodic inspection of bridges to identify damages can aid in their quick remediation. Likewise, identifying components can provide context for damage assessment and help gauge a bridge's state of interaction with its surroundings. Current inspection techniques rely on manual site visits, which can be time-consuming and costly. More recently, robotic inspection assisted by autonomous data analytics based on Computer Vision (CV) and Artificial Intelligence (AI) has been viewed as a suitable alternative to manual inspection because of its efficiency and accuracy. To aid research in this avenue, this study performs a comparative assessment of different architectures, loss functions, and ensembling strategies for the autonomous segmentation of bridge components and damages. The experiments lead to several interesting discoveries. Nested Reg-UNet architecture is found to outperform five other state-of-the-art architectures in both damage and component segmentation tasks. The architecture is built by combining a Nested UNet style dense configuration with a pretrained RegNet encoder. In terms of the mean Intersection over Union (mIoU) metric, the Nested Reg-UNet architecture provides an improvement of 2.86% on the damage segmentation task and 1.66% on the component segmentation task compared to the state-of-the-art UNet architecture. Furthermore, it is demonstrated that incorporating the Lovasz-Softmax loss function to counter class imbalance can boost performance by 3.44% in the component segmentation task over the most employed alternative, weighted Cross Entropy (wCE). Finally, weighted softmax ensembling is found to be quite effective when used synchronously with the Nested Reg-UNet architecture by providing mIoU improvement of 0.74% in the component segmentation task and 1.14% in the damage segmentation task over a single-architecture baseline. Overall, the best mIoU of 92.50% for the component segmentation task and 84.19% for the damage segmentation task validate the feasibility of these techniques for autonomous bridge component and damage segmentation using RGB images.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.9
• /
• pp.37-47
• /
• 2021

In this paper, we propose a new deep learning model for tongue segmentation with improved accuracy compared to the existing model by diversifying the receptive field in the U-net. Methods such as parallel convolution, dilated convolution, and constant channel increase were used to diversify the receptive field. For the proposed deep learning model, a tongue region segmentation experiment was performed on two test datasets. The training image and the test image are similar in TestSet1 and they are not in TestSet2. Experimental results show that segmentation performance improved as the receptive field was diversified. The mIoU value of the proposed method was 98.14% for TestSet1 and 91.90% for TestSet2 which was higher than the result of existing models such as U-net, DeepTongue, and TongueNet.

• Smart Media Journal
• /
• v.11 no.4
• /
• pp.46-55
• /
• 2022

It is very important to first detect and remove defective fruits with scratches or bruised areas in the automatic fruit quality screening system. This paper proposes a method of detecting defective areas in fruits using the latest artificial intelligence technology, the Yolo V4 deep learning model in order to overcome the limitations of the method of detecting fruit's defective areas using the existing image processing techniques. In this study, a total of 2,400 defective fruits, including 1,000 defective apples and 1,400 defective fruits with scratch or decayed areas, were learned using the Yolo V4 deep learning model and experiments were conducted to detect defective areas. As a result of the performance test, the precision of apples is 0.80, recall is 0.76, IoU is 69.92% and mAP is 65.27%. The precision of pears is 0.86, recall is 0.81, IoU is 70.54% and mAP is 68.75%. The method proposed in this study can dramatically improve the performance of the existing automatic fruit quality screening system by accurately selecting fruits with defective areas in real time rather than using the existing image processing techniques.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.153-166
• /
• 2022

Identifying fine cracks in steel bridge facilities is a challenging task of structural health monitoring (SHM). This study proposed an end-to-end crack image segmentation framework based on a one-step Convolutional Neural Network (CNN) for pixel-level object recognition with high accuracy. To particularly address the challenges arising from small object detection in complex background, efforts were made in loss function selection aiming at sample imbalance and module modification in order to improve the generalization ability on complicated images. Specifically, loss functions were compared among alternatives including the Binary Cross Entropy (BCE), Focal, Tversky and Dice loss, with the last three specialized for biased sample distribution. Structural modifications with dilated convolution, Spatial Pyramid Pooling (SPP) and Feature Pyramid Network (FPN) were also performed to form a new backbone termed CrackDet. Models of various loss functions and feature extraction modules were trained on crack images and tested on full-scale images collected on steel box girders. The CNN model incorporated the classic U-Net as its backbone, and Dice loss as its loss function achieved the highest mean Intersection-over-Union (mIoU) of 0.7571 on full-scale pictures. In contrast, the best performance on cropped crack images was achieved by integrating CrackDet with Dice loss at a mIoU of 0.7670.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.6
• /
• pp.392-399
• /
• 2023

This paper proposes a deep learning model to determine the region and depth of cabbage cores for robotic automation of the cabbage core removal process during the kimchi manufacturing process. In addition, rather than predicting the depth of the measured cabbage, a model was presented that simultaneously detects and classifies the area by converting it into a discrete class. For deep learning model learning and verification, RGB images of the harvested cabbage 522 were obtained. The core region and depth labeling and data augmentation techniques from the acquired images was processed. MAP, IoU, acuity, sensitivity, specificity, and F1-score were selected to evaluate the performance of the proposed YOLO-v4 deep learning model-based cabbage core area detection and classification model. As a result, the mAP and IoU values were 0.97 and 0.91, respectively, and the acuity and F1-score values were 96.2% and 95.5% for depth classification, respectively. Through the results of this study, it was confirmed that the depth information of cabbage can be classified, and that it can be used in the development of a robot-automation system for the cabbage core removal process in the future.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.8
• /
• pp.1996-2015
• /
• 2023

High-definition (HD) maps can provide precise road information that enables an autonomous driving system to effectively navigate a vehicle. Recent research has focused on leveraging semantic segmentation to achieve automatic annotation of HD maps. However, the existing methods suffer from low recognition accuracy in automatic driving scenarios, leading to inefficient annotation processes. In this paper, we propose a novel semantic segmentation method for automatic HD map annotation. Our approach introduces a new encoder, known as the convolutional transformer hybrid encoder, to enhance the model's feature extraction capabilities. Additionally, we propose a multi-level fusion module that enables the model to aggregate different levels of detail and semantic information. Furthermore, we present a novel decoupled boundary joint decoder to improve the model's ability to handle the boundary between categories. To evaluate our method, we conducted experiments using the Bird's Eye View point cloud images dataset and Cityscapes dataset. Comparative analysis against stateof-the-art methods demonstrates that our model achieves the highest performance. Specifically, our model achieves an mIoU of 56.26%, surpassing the results of SegFormer with an mIoU of 1.47%. This innovative promises to significantly enhance the efficiency of HD map automatic annotation.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.3
• /
• pp.17-22
• /
• 2023

In this paper, we propose a black ice detection system using Convolutional Neural Networks (CNNs). Black ice poses a serious threat to road safety, particularly during winter conditions. To overcome this problem, we introduce a CNN-based architecture for real-time black ice detection with an encoder-decoder network, specifically designed for real-time black ice detection using thermal images. To train the network, we establish a specialized experimental platform to capture thermal images of various black ice formations on diverse road surfaces, including cement and asphalt. This enables us to curate a comprehensive dataset of thermal road black ice images for a training and evaluation purpose. Additionally, in order to enhance the accuracy of black ice detection, we propose a multi-scale dilation convolution feature fusion (MsDC-FF) technique. This proposed technique dynamically adjusts the dilation ratios based on the input image's resolution, improving the network's ability to capture fine-grained details. Experimental results demonstrate the superior performance of our proposed network model compared to conventional image segmentation models. Our model achieved an mIoU of 95.93%, while LinkNet achieved an mIoU of 95.39%. Therefore, it is concluded that the proposed model in this paper could offer a promising solution for real-time black ice detection, thereby enhancing road safety during winter conditions.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.538-540
• /
• 2021

In this paper, we proposed a smart target detection system that detects and recognizes a designated target to provide relative motion information when performing a target detection mission of a drone. The proposed system focused on developing an algorithm that can secure adequate accuracy (i.e. mAP, IoU) and high real-time at the same time. The proposed system showed an accuracy of close to 1.0 after 100k learning of the Google Inception V2 deep learning model, and the inference speed was about 60-80[Hz] when using a high-performance laptop based on the real-time performance Nvidia GTX 2070 Max-Q. The proposed smart target detection system will be operated like a drone and will be helpful in successfully performing surveillance and reconnaissance missions by automatically recognizing the target using computer image processing and following the target.

• Smart Structures and Systems
• /
• v.31 no.4
• /
• pp.383-392
• /
• 2023

Post-earthquake building condition assessment is crucial for subsequent rescue and remediation and can be automated by emerging computer vision and deep learning technologies. This study is based on an endeavour for the 2nd International Competition of Structural Health Monitoring (IC-SHM 2021). The task package includes five image segmentation objectives - defects (crack/spall/rebar exposure), structural component, and damage state. The structural component and damage state tasks are identified as the priority that can form actionable decisions. A multi-task Convolutional Neural Network (CNN) is proposed to conduct the two major tasks simultaneously. The rest 3 sub-tasks (spall/crack/rebar exposure) were incorporated as auxiliary tasks. By synchronously learning defect information (spall/crack/rebar exposure), the multi-task CNN model outperforms the counterpart single-task models in recognizing structural components and estimating damage states. Particularly, the pixel-level damage state estimation witnesses a mIoU (mean intersection over union) improvement from 0.5855 to 0.6374. For the defect detection tasks, rebar exposure is omitted due to the extremely biased sample distribution. The segmentations of crack and spall are automated by single-task U-Net but with extra efforts to resample the provided data. The segmentation of small objects (spall and crack) benefits from the resampling method, with a substantial IoU increment of nearly 10%.