• Journal of Sensor Science and Technology
• /
• v.33 no.3
• /
• pp.119-124
• /
• 2024

Effective fire sensing is important to protect lives and property from the disaster. In this paper, we present an intelligent visual sensing method for detecting fires based on machine learning techniques. The proposed method involves a two-step process. In the first step, fire and non-fire images are used to train a convolutional neural network (CNN), and in the next step, feature vectors consisting of 256 values obtained from the CNN are used for the learning of a support vector machine (SVM). Linear and nonlinear SVMs with different parameters are intensively tested. We found that the proposed hybrid method using an SVM with a linear kernel effectively increased the recall rate of fire image detection without compromising detection accuracy when an imbalanced dataset was used for learning. This is a major contribution of this study because recall is important, particularly in the sensing of disaster situations such as fires. In our experiments, the proposed system exhibited an accuracy of 96.9% and a recall rate of 92.9% for test image data.

• International Journal of Highway Engineering
• /
• v.19 no.5
• /
• pp.69-75
• /
• 2017

PURPOSES : The purpose of this study is to evaluate different types of Ground Penetrating Radar (GPR) testing for characterizing the road cavity detection. The impulse and step-frequency-type GPR tests were conducted on a full-scale testbed with an artificial void installation. After analyzing the response signals of GPR tests for detecting the road cavity, the characteristics of each GPR response was evaluated for a suitable selection of GPR tests. METHODS : Two different types of GPR tests were performed to estimate the limitation and accuracy for detecting the cavities underneath the asphalt pavement. The GPR signal responses were obtained from the testbed with different cavity sizes and depths. The detection limitation was identified by a signal penetration depth at a given cavity for impulse and step-frequency-type GPR testing. The unique signal characteristics was also observed at cavity sections. RESULTS : The impulse-type GPR detected the 500-mm length of cavity at a depth of 1.0 m, and the step-frequency-type GPR detected the cavity up to 1.5 m. This indicates that the detection capacity of the step-frequency type is better than the impulse type. The step-frequency GPR testing also can reflect the howling phenomena that can more accurately determine the cavity. CONCLUSIONS :It is found from this study that the step-frequency GPR testing is more suitable for the road cavity detection of asphalt pavement. The use of step-frequency GPR testing shows a distinct image at the cavity occurrences.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.11 no.5
• /
• pp.277-285
• /
• 2016

Moving object detection system has been an emerging research field in various advanced driver assistance systems (ADAS) and surveillance system. In this paper, we propose two optical flow based moving object detection methods at dynamic scenes. Both proposed methods consist of three successive steps; pre-processing, foreground segmentation, and post-processing steps. Two proposed methods have the same pre-processing and post-processing steps, but different foreground segmentation step. Pre-processing calculates mainly optical flow map of which each pixel has the amplitude of motion vector. Dense optical flows are estimated by using Farneback technique, and the amplitude of the motion normalized into the range from 0 to 255 is assigned to each pixel of optical flow map. In the foreground segmentation step, moving object and background are classified by using the optical flow map. Here, we proposed two algorithms. One is Gaussian mixture model (GMM) based background subtraction, which is applied on optical map. Another is adaptive thresholding based foreground segmentation, which classifies each pixel into object and background by updating threshold value column by column. Through the simulations, we show that both optical flow based methods can achieve good enough object detection performances in dynamic scenes.

• 제어로봇시스템학회:학술대회논문집
• /
• 1987.10b
• /
• pp.566-569
• /
• 1987

The primary difficulties encountered in the use of step motors are underdamped response when stopping at a specified position and dynamic instability during high-speed slewing. This paper proposes a speed and position detection scheme using the back EMF generated by the rotating permanent magnet field of a two-phase 1.8.deg. hybrid step motor, and presents its application to the state-variable feedback control of the hybrid step motor. All simulation results in a single step response show that the hybrid step motor performances such as peak overshoot and settling time are greatly improved.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.10
• /
• pp.4661-4680
• /
• 2016

Cognitive radio (CR) technology is an effective solution to the spectrum scarcity issue. Collaborative spectrum sensing is known as a promising technique to improve the performance of spectrum sensing in cognitive radio networks (CRNs). However, collaborative spectrum sensing is vulnerable to spectrum data falsification (SSDF) attack, where malicious users (MUs) may send false sensing data to mislead other secondary users (SUs) to make an incorrect decision about primary user (PUs) activity, which is one of the key adversaries to the performance of CRNs. In this paper, we propose a coalition based malicious users detection (CMD) algorithm to detect the malicious user in CRNs. The proposed CMD algorithm can efficiently detect MUs base on the Geary'C theory and be modeled as a coalition formation game. Specifically, SSDF attack is one of the key issues to affect the resource allocation process. Focusing on the security issues, in this paper, we analyze the power allocation problem with MUs, and propose MUs detection based power allocation (MPA) algorithm. The MPA algorithm is divided into two steps: the MUs detection step and the optimal power allocation step. Firstly, in the MUs detection step, by the CMD algorithm we can obtain the MUs detection probability and the energy consumption of MUs detection. Secondly, in the optimal power allocation step, we use the Lagrange dual decomposition method to obtain the optimal transmission power of each SU and achieve the maximum utility of the whole CRN. Numerical simulation results show that the proposed CMD and MPA scheme can achieve a considerable performance improvement in MUs detection and power allocation.

• ETRI Journal
• /
• v.37 no.3
• /
• pp.551-561
• /
• 2015

We propose a novel multiple-object tracking algorithm for real-time intelligent video surveillance. We adopt particle filtering as our tracking framework. Background modeling and subtraction are used to generate a region of interest. A two-step pedestrian detection is employed to reduce the computation time of the algorithm, and an iterative particle repropagation method is proposed to enhance its tracking accuracy. A matching score for greedy data association is proposed to assign the detection results of the two-step pedestrian detector to trackers. Various experimental results demonstrate that the proposed algorithm tracks multiple objects accurately and precisely in real time.

• Journal of Information Processing Systems
• /
• v.14 no.1
• /
• pp.150-161
• /
• 2018

We propose a novel deep neural network model for detecting human activities in untrimmed videos. The process of human activity detection in a video involves two steps: a step to extract features that are effective in recognizing human activities in a long untrimmed video, followed by a step to detect human activities from those extracted features. To extract the rich features from video segments that could express unique patterns for each activity, we employ two different convolutional neural network models, C3D and I-ResNet. For detecting human activities from the sequence of extracted feature vectors, we use BLSTM, a bi-directional recurrent neural network model. By conducting experiments with ActivityNet 200, a large-scale benchmark dataset, we show the high performance of the proposed DeepAct model.

• Smart Media Journal
• /
• v.2 no.4
• /
• pp.20-26
• /
• 2013

This paper presents an approach to Mongolian car plate recognition using artificial neural network. Our proposed method consists of two steps: detection and recognition. In detection step, we implement Flood fill algorithm. In recognition step we proceed to segment the plate for each Cyrillic character, and use an Artificial Neural Network (ANN) machine - learning algorithm to recognize the character. We have learned the theory of ANN and implemented it without using any library. A total of 150 vehicles images obtained from community entrance gates have been tested. The recognition algorithm shows an accuracy rate of 89.75%.

• Steel and Composite Structures
• /
• v.25 no.4
• /
• pp.485-496
• /
• 2017

In this paper, damage detection has been introduced as an optimization problem and a two-step method has been proposed that can detect the location and severity of damage in truss structures precisely and reduce the volume of computations considerably. In the first step, using the residual force vector concept, the suspected damaged members are detected which will result in a reduction in the number of variables and hence a decrease in the search space dimensions. In the second step, the precise location and severity of damage in the members are identified using the genetic algorithm and the results of the first step. Considering the reduced search space, the algorithm can find the optimal points (i.e. the solution for the damage detection problem) with less computation cost. In this step, the Efficient Correlation Based Index (ECBI), that considers the structure's first few frequencies in both damaged and healthy states, is used as the objective function and some examples have been provided to check the efficiency of the proposed method; results have shown that the method is innovatively capable of detecting damage in truss structures.

• Journal of Institute of Control, Robotics and Systems
• /
• v.2 no.4
• /
• pp.337-343
• /
• 1996

In this paper, we investigate the defect detection by making use of pre-made reference image data and classify the defects by using the artificial neural network. The approach is composed of three main parts. The first step consists of a proper generation of two reference image data by using a low level morphological technique. The second step proceeds by performing three times logical bit operations between two ready-made reference images and just captured image to be tested. This results in defects image only. In the third step, by extracting four features from each detected defect, followed by assigning them into the input nodes of an already trained artificial neural network we can obtain a defect class corresponding to the features. All of the image data are formed in a bit level for the reduction of data size as well as time saving. Experimental results show that proposed algorithms are found to be effective for flexible defect detection, robust classification, and high speed process by adopting a simple logic operation.