• Journal of the Korean Society for Precision Engineering
• /
• v.16 no.9
• /
• pp.25-34
• /
• 1999

In spite of the advance of 3D solid modeling technology, there are some distinct areas where 2D CAD S/W are still dominant, and more competent comparing with 3D CAD S/W. For example, in the manufacturing of 2D-shaped electrical parts, most related manufacturing tools have 2D geometric features by nature, and 3D solid models applied to these parts have substantial overheads. Nevertheless, most 2D CAD S/W have no topological inquiry services because they have no such information on their geometrical database inherently. Thus, it is needed to extract such information from 2D CAD database for developing more advanced application such as automated drafting/design S/W. In this paper, the extraction of topological information from 2D CAD has been performed in general way using concept of B-rep. A general extraction algorithm, data structure and meta file format for 2D topological object have been developed and successfully applied to the development of the automated lead frame die design system in Samsung Aerospace. it is also possible to provide a flexible, powerful topology-oriented functionality on any common 2D CAD S/W.

• Journal of the Korean Society for Precision Engineering
• /
• v.25 no.5
• /
• pp.82-88
• /
• 2008

This paper proposes a novel method of 2 dimensional shape display. Shape displays allow us to feel tile actual volume of the object, unlike conventional 2D visual displays of 3D objects. The proposed method employs a wireframe structure to present 2D or 3D objects. The wireframe is composed of small units driven by shape memory alloy (SMA) actuators. The drive unit is analogous to the agonist-antagonist system of animal musculoskeletal systems, where the SMA actuators serve as agonist and antagonist muscles. The force in the SMA actuator is controlled by electrical current. The drive unit is equipped with the locking mechanism so that it can sustain the external force exerted by the user as well as the own weight of the wireframe structure. By controlling the current into the SMA actuator and locking mechanism, we can control the angle of the drive unit. A chain of drive units enables presentation of 2 dimensional objects. 3 dimensional presentations are possible by collecting the chains of drive units.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.26 no.2
• /
• pp.129-152
• /
• 2024

In the application of deep learning object detection via CCTV in tunnels, a large number of false positive detections occur due to the poor environmental conditions of tunnels, such as low illumination and severe perspective effect. This problem directly impacts the reliability of the tunnel CCTV-based accident detection system reliant on object detection performance. Hence, it is necessary to reduce the number of false positive detections while also enhancing the number of true positive detections. Based on a deep learning object detection model, this paper proposes a false positive data training method that not only reduces false positives but also improves true positive detection performance through retraining of false positive data. This paper's false positive data training method is based on the following steps: initial training of a training dataset - inference of a validation dataset - correction of false positive data and dataset composition - addition to the training dataset and retraining. In this paper, experiments were conducted to verify the performance of this method. First, the optimal hyperparameters of the deep learning object detection model to be applied in this experiment were determined through previous experiments. Then, in this experiment, training image format was determined, and experiments were conducted sequentially to check the long-term performance improvement through retraining of repeated false detection datasets. As a result, in the first experiment, it was found that the inclusion of the background in the inferred image was more advantageous for object detection performance than the removal of the background excluding the object. In the second experiment, it was found that retraining by accumulating false positives from each level of retraining was more advantageous than retraining independently for each level of retraining in terms of continuous improvement of object detection performance. After retraining the false positive data with the method determined in the two experiments, the car object class showed excellent inference performance with an AP value of 0.95 or higher after the first retraining, and by the fifth retraining, the inference performance was improved by about 1.06 times compared to the initial inference. And the person object class continued to improve its inference performance as retraining progressed, and by the 18th retraining, it showed that it could self-improve its inference performance by more than 2.3 times compared to the initial inference.

• Proceedings of the KSME Conference
• /
• 2007.05a
• /
• pp.1505-1510
• /
• 2007

Recently, efficient manufacturing of high precision is an important issue in modern industry as more variety of industrial products is being designed with compound surfaces. Rapid CAD data generation can be possible based on a non-contact type CMM of object through the use of reverse engineering. However, some registration to match the data measured from various directions into a common coordinate system is required. Also, the error can happen if it uses the conventional method to large product of thin thickness. So it is necessary to develop a new method, which was designed for the registration of large and thin products. Additionally, an algorithm to pick up coordinates for the newly designed method was proposed.

• Tribology and Lubricants
• /
• v.23 no.2
• /
• pp.56-60
• /
• 2007

The main object of this study is to compare the results that have been concluded by the experiment and to estimate the temperature rise that can cause the contacting surface to be damaged. The former studies are based on the Blok and Jaeger formula. By these formulas we assume that two of the contacted objects are a kind of semi-infinite solid and with this assumption we can make a temperature analysis. But this method doesn't consider lubrication conditions and the calculation time requires a lot of time in that we have to face many difficulties in measuring the actual temperature rise. In this study we combines the semi-infinite solid method and the finite volume method to analyze the temperature of the contacting surface. And we measure temperature rise of the contact surface by dynamic thermocouple.

• International Journal of High-Rise Buildings
• /
• v.9 no.4
• /
• pp.335-342
• /
• 2020

This article comes up with a new method which is based on the visual characteristic of the objects and machine learning technology to achieve semi-automated recognition of the personnel, machine & materials of the construction sites. Balancing the real-time performance and accuracy, using Faster RCNN (Faster Region-based Convolutional Neural Networks) with transfer learning method appears to be a rational choice. After fine-tuning an ImageNet pre-trained Faster RCNN and testing with it, the result shows that the precision ratio (mAP) has so far reached 67.62%, while the recall ratio (AR) has reached 56.23%. In other word, this recognizing method has achieved rational performance. Further inference with the video of the construction of Huoshenshan Hospital also indicates preliminary success.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.6 no.2
• /
• pp.10-18
• /
• 1988

In the three-dimensional analysis and deformation analysis of large structures, efficient is the use of the multiple method of close range photogrammetry which approaches the object distance. This study analyzes the influence of errors according to the overlap, the control points, and the object distance, to solve the problems which are raised in the multiple method. A wall-board, 7 meters by 3 meters, was used as a test field on which a total of 225 unknown points were equally disposed. The photographs with changing the overlap and object distance were taken by P-31 camera system. a total of 143 negatives are used in this study for computing 3-dimensional coordinates and its standard errors, and bundle adjustment of strips and blocks developed with on-line system is applied. In case of decreasing the number of control points, simulation error increases but actual error decreases and increases again. Due to the changed of object distances Z error represents largely compared to X, Y error, but good results in Z can be obtained by increasing the redundancy. And simulation error or actual error shows best results at the endlap of about 70%. To sum up this study, approprate arrangement of control points and overlap is meaningful, and multiple method by short object distance will be widely used to precision and deformation analysis of critical structures.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.11
• /
• pp.153-161
• /
• 2024

In this paper, we address the limitation of single-channel infrared (IR) images, which are difficult to directly apply to RGB-based detection models. Previously, a single channel was often replicated into three channels; however, this approach may limit detection performance due to information redundancy. To overcome this limitation, we propose a method that replicates the single-channel IR image into three channels, with each channel processed using different preprocessing techniques, such as CLAHE (Contrast Limited Adaptive Histogram Equalization), Laplacian Filter, and Top-hat transform, to improve detection performance. In this study, we utilized the RT-DETRv2 detection model and the Anti-UAV300 dataset, using IR images sampled at 10-frame intervals for our experiments. By evaluating the effects of each preprocessing technique and deriving the optimal configuration, our method achieved a 2.2% improvement in mean Average Precision (mAP) over conventional methods. This confirms that our method enhances performance over simple replication, presenting a novel approach to improving object detection performance in IR imaging, with promising applications across various fields, particularly in disaster situations where infrared cameras are utilized, as well as in nighttime surveillance and reconnaissance.

• Journal of Information Processing Systems
• /
• v.13 no.5
• /
• pp.1372-1381
• /
• 2017

In this paper, a texture feature extraction method using local energy and local correlation of Gabor transformed images is proposed and applied to an image retrieval system. The Gabor wavelet is known to be similar to the response of the human visual system. The outputs of the Gabor transformation are robust to variants of object size and illumination. Due to such advantages, it has been actively studied in various fields such as image retrieval, classification, analysis, etc. In this paper, in order to fully exploit the superior aspects of Gabor wavelet, local energy and local correlation features are extracted from Gabor transformed images and then applied to an image retrieval system. Some experiments are conducted to compare the performance of the proposed method with those of the conventional Gabor method and the popular rotation-invariant uniform local binary pattern (RULBP) method in terms of precision vs recall. The Mahalanobis distance is used to measure the similarity between a query image and a database (DB) image. Experimental results for Corel DB and VisTex DB show that the proposed method is superior to the conventional Gabor method. The proposed method also yields precision and recall 6.58% and 3.66% higher on average in Corel DB, respectively, and 4.87% and 3.37% higher on average in VisTex DB, respectively, than the popular RULBP method.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.8
• /
• pp.369-376
• /
• 2016

Among various vehicle extraction techniques, OBPCA (Object-Based Point Cloud Analysis) calculates features quickly by coarse-grained rectangles from top-view of the vehicle candidates. However, it uses only a top-view rectangle to detect a vehicle. Thus, it is hard to extract rectangular objects with similar size. For this reason, accuracy issue has raised on the OBPCA method which influences on DEM generation and traffic monitoring tasks. In this paper, we propose a novel method which uses the most distinguishing vertical elevations to calculate additional features. Our proposed method uses same features with top-view, determines new thresholds, and decides whether the candidate is vehicle or not. We compared the accuracy and execution time between original OBPCA and the proposed one. The experiment result shows that our method produces 6.61% increase of precision and 13.96% decrease of false positive rate despite with marginal increase of execution time. We can see that the proposed method can reduce misclassification.