• Smart Structures and Systems
• /
• v.33 no.2
• /
• pp.105-118
• /
• 2024

In order to realize tiny bridge crack discovery by UAV-based machine vision, a novel method combining deep learning and tensor voting is proposed. Firstly, the grid images of crack are detected and descripted based on SE-ResNet50 to generate feature points. Then, the probability significance map of crack image is calculated by tensor voting with feature points, which can define the direction and region of crack. Further, the crack detection anchor box is formed by non-maximum suppression from the probability significance map, which can improve the robustness of tiny crack detection. Finally, a case study is carried out to demonstrate the effectiveness of the proposed method in the Xiangjiang-River bridge inspection. Compared with the original tensor voting algorithm, the proposed method has higher accuracy in the situation of only 1-2 pixels width crack and the existence of edge blur, crack discontinuity, which is suitable for UAV-based bridge crack discovery.

• Journal of Institute of Control, Robotics and Systems
• /
• v.6 no.6
• /
• pp.486-491
• /
• 2000

The dicing process cuts a semiconductor wafer to lengthwise and crosswise direction by using a rotating circular diamond blade. But inferior goods are made under the influence of several parameters in dicing such as blade, wafer, cutting water and cutting conditions. This paper describes a monitoring algorithm using neural network in order to find out an instant of vibration signal change when bad dicing appears. The algorithm is composed of two steps: feature extraction and decision. In the feature extraction, five features processed from vibration signal which is acquired by accelerometer attached on blade head are proposed. In the decision, back-propagation neural network is adopted to classify the dicing process into normal and abnormal dicing, and normal and damaged blade. Experiments have been performed for GaAs semiconductor wafer in the case of normal/abnormal dicing and normal/damaged blade. Based upon observation of the experimental results, the proposed scheme shown has a good accuracy of classification performance by which the inferior goods decreased from 35.2% to 6.5%.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.8
• /
• pp.1693-1698
• /
• 2009

This paper is a research on the feature information using direction information and adjusting constant w for the dynamic signature verification. We could improved processing time and reduce signature database without the increase of error rate. We could confirmed these results by using statistical method T-test.

• Journal of Korea Multimedia Society
• /
• v.20 no.12
• /
• pp.2009-2015
• /
• 2017

A two-dimensional lens is required to reproduce both the horizontal and vertical parallax through an autostereoscopic 3D display. Among the two-dimensional lenses, a hexagonal micro lens array (MLA) having good optical efficiency is mainly used. However, the hexagonal MLA has complex geometric features. The first feature is that the lens cells are zigzagged in the vertical direction, which should be reflected in the view number calculation for each sub-pixel. The second feature is that the four sides of a hexagonal lens cell are tilted, requiring a more careful view index assignment to the lens cell. In this paper, we propose a sub-pixel multiplexing scheme suitable for the features of the hexagonal MLA. We also propose a view-overlay algorithm based on a two-dimensional lens and compare subjective image quality with existing view-selection through autostereoscopic 3D display implementation.

• International Journal of Contents
• /
• v.4 no.2
• /
• pp.24-28
• /
• 2008

In this paper, we propose a new approach for content-based video retrieval using non-parametric based motion classification in the shot-based video indexing structure. Our system proposed in this paper has supported the real-time video retrieval using spatio-temporal feature comparison by measuring the similarity between visual features and between motion features, respectively, after extracting representative frame and non-parametric motion information from shot-based video clips segmented by scene change detection method. The extraction of non-parametric based motion features, after the normalized motion vectors are created from an MPEG-compressed stream, is effectively fulfilled by discretizing each normalized motion vector into various angle bins, and by considering the mean, variance, and direction of motion vectors in these bins. To obtain visual feature in representative frame, we use the edge-based spatial descriptor. Experimental results show that our approach is superior to conventional methods with regard to the performance for video indexing and retrieval.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.49 no.6
• /
• pp.307-315
• /
• 2000

For the autonomy of a mobile robot it is first needed to know its position and orientation. Various methods of estimating the position of a robot have been developed. However, it is still difficult to localize the robot without any initial position or orientation. In this paper we present the method how to make the colored map and how to calculate the position and direction of a robot using the angle data of an omnidirectional image. The wall of the map is rendered with the corresponding color images and the color histograms of images and the coordinates of feature points are stored in the map. Then a mobile robot gets the color omnidirectional image at arbitrary position and orientation, segments it and recognizes objects by multiple color indexing. Using the information of recognized objects robot can have enough feature points and localize itself.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.4
• /
• pp.129-137
• /
• 1994

In this paper, a new texture feature extraction method for texture image classification is proposed. The proposed method is a modified texture spectrum method. It uses local weighting averaged texture unit, that is, the neighbor pixels are weithted and averaged in 4-direction and the calculated values are compared with center pixel to find texture units. The proposed method has only 81 texture units and these units are really good features for texture classification. The proposed method is applied to vegetable images and Blodatz album images and compared with several conventional methods for the feature extraction time and the recognition rate.

• Journal of the Korean Society for Precision Engineering
• /
• v.22 no.10 s.175
• /
• pp.107-113
• /
• 2005

This paper presents a method to reduce errors in registration, which is used in transformation coordinate system of the multiple measuring data. In general, the ICP algorithms and feature-based approaches are used for registration. In order to measure wrap-around object, it is necessary to change the scanning direction or set-up of the object. A fixture is made to reduce registration errors caused by inaccurate center point of tooling balls, providing the more accurate registration method. And, the motorized fixture controls rotation and tilting to get precise the measuring data and registration. The proposed motorized fixture and registration method have advantages in accurate registration and precise measurement, compared with the conventional methods.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.3 no.1
• /
• pp.100-104
• /
• 2003

The SVDD(support vector data description) is one of the most well-known one-class support vector learning methods, in which one tries the strategy of utilizing balls defined on the kernel feature space in order to distinguish a set of normal data from all other possible abnormal objects. The major concern of this paper is to consider the problem of modifying the SVDD into the direction of utilizing ellipsoids instead of balls in order to enable better classification performance. After a brief review about the original SVDD method, this paper establishes a new method utilizing ellipsoids in feature space, and presents a solution in the form of SDP(semi-definite programming) which is an optimization problem based on linear matrix inequalities.

• Journal of Institute of Control, Robotics and Systems
• /
• v.12 no.5
• /
• pp.473-479
• /
• 2006

A global path planning algorithm using modified self-organizing feature map(SOFM) which is a method among a number of neural network is presented. The SOFM uses a randomized small valued initial weight vectors, selects the neuron whose weight vector best matches input as the winning neuron, and trains the weight vectors such that neurons within the activity bubble are move toward the input vector. On the other hand, the modified method in this research uses a predetermined initial weight vectors of the 2-dimensional mesh, gives the systematic input vector whose position best matches obstacles, and trains the weight vectors such that neurons within the activity bubble are move toward the opposite direction of input vector. According to simulation results one can conclude that the modified neural network is useful tool for the global path planning problem of a mobile robot.