• Journal of the Korean institute of surface engineering
• /
• v.55 no.6
• /
• pp.398-402
• /
• 2022

Vacuum Insulation Panel (VIP) is an high energy efficient insulation system that facilitate slim but high insulation performance, based on based on a porous core material evacuated and encapsulated in a multi-barrier envelope. Although VIP has been on the market for decades now, it wasn't until recently that efforts have been initiated to propose a standard on aging testing. One of the issues regarding VIP is its durability and aging due to pressure and moisture dependent increase of the initial low thermal conductivity with time. It is hard to visually determine at an early stage. Recently, a method of analyzing the damage on the a material surface by applying image processing technology has been widely used. These techniques provide fast and accurate data with a non-destructive way. In this study, the surface VIP images were analyzed using the Harris corner detection algorithm. As a result, 171,333 corner points in the normal packaging were detected, whereas 32,895 of the defective packaging, which were less than the normal packaging. were detected. These results are considered to provide meaningful information for the determination of VIP condition.

• Journal of Information Processing Systems
• /
• v.18 no.5
• /
• pp.677-687
• /
• 2022

We propose a detection algorithm based on tree-structured convolutional neural networks (TsCNNs) that finds pornography, propaganda, or other inappropriate content on a social media network. The algorithm sequentially applies the typical convolutional neural network (CNN) algorithm in a tree-like structure to minimize classification errors in similar classes, and thus improves accuracy. We implemented the detection system and conducted experiments on a data set comprised of 6 ordinary classes and 11 inappropriate classes collected from the Korean military social network. Each model of the proposed algorithm was trained, and the performance was then evaluated according to the images and videos identified. Experimental results with 20,005 new images showed that the overall accuracy in image identification achieved a high-performance level of 99.51%, and the effectiveness of the algorithm reduced identification errors by the typical CNN algorithm by 64.87 %. By reducing false alarms in video identification from the domain, the TsCNNs achieved optimal performance of 98.11% when using 10 minutes frame-sampling intervals. This indicates that classification through proper sampling contributes to the reduction of computational burden and false alarms.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.5
• /
• pp.919-926
• /
• 2023

In this paper, we implemented the surveillance camera system capable of day and night shooting. To this end, it is designed to capture clear images even at night using a CMOS image sensor as well as an IR-LED. In addition, a relatively simple motion detection algorithm was proposed through color model separation. Motions can be detected by extracting only the H channel from the color model, dividing the image into blocks, and then applying the block matching method using the average color value between consecutive frames. When motions are detected during filming, an alarm sounds automatically and a day and night motion detection system is implemented that can capture and save the event screen to a PC.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.8
• /
• pp.3981-4004
• /
• 2019

This paper proposes a novel method for locating objects in real space from a single remote image and measuring actual distances between them by automatic detection and perspective transformation. The dimensions of the real space are known in advance. First, the corner points of the interested region are detected from an image using deep learning. Then, based on the corner points, the region of interest (ROI) is extracted and made proportional to real space by applying warp-perspective transformation. Finally, the objects are detected and mapped to the real-world location. Removing distortion from the image using camera calibration improves the accuracy in most of the cases. The deep learning framework Darknet is used for detection, and necessary modifications are made to integrate perspective transformation, camera calibration, un-distortion, etc. Experiments are performed with two types of cameras, one with barrel and the other with pincushion distortions. The results show that the difference between calculated distances and measured on real space with measurement tapes are very small; approximately 1 cm on an average. Furthermore, automatic corner detection allows the system to be used with any type of camera that has a fixed pose or in motion; using more points significantly enhances the accuracy of real-world mapping even without camera calibration. Perspective transformation also increases the object detection efficiency by making unified sizes of all objects.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.10 no.1
• /
• pp.13-23
• /
• 2015

Recently, a few image-processing based mobile urine testers have actively been studied since the urine-analysis result can be available to the user in real time immediately after the test is done. However, the accuracy of test result can be severely degraded due to variable illumination environments and a variety of manners to capture the image with a camera embedded in the smartphone according to different users. This paper proposes the Check4Urine system, a novel smartphone-based portable urine-analysis tester and provides three techniques to improve such a performance degradation problem robust to various test environments and disturbances, which are the compensation algorithm to correct the varying illumination effect, an urine strip detection algorithm robust to edge loss of the object image, and the color decision algorithm based on the pre-processed reference table. Experimental results show that the proposed Check4Urine system increases the accuracy of urine-analysis by 20-50% at various test conditions, compared with the existing image-processing based mobile urine tester.

• The Journal of Korea Robotics Society
• /
• v.9 no.3
• /
• pp.140-146
• /
• 2014

This paper introduces a dual-arm robot painter system which is capable of sketching a camera-captured image with short line segments. To express various curved edges in the image by combining line segments, we first apply edge detection algorithm to the entire image, split the edged image into small boxed pieces, and then apply Hough Transformation to each piece so that the edges inside the piece can be approximated with short line segments. To draw the picture within a reasonable time, we designed a simple dual-arm robot system and controlled both arms concurrently according to linear interpolation algorithm. From the experiments, we could verify that simple linear motions can describe various images effectively with a unique brush style.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.11 no.3
• /
• pp.163-171
• /
• 2016

Lane detection is a widely researched topic. Although simple road detection is easily achieved by previous methods, lane detection becomes very difficult in several complex cases involving noisy edges. To address this, we use a Convolution neural network (CNN) for image enhancement. CNN is a deep learning method that has been very successfully applied in object detection and recognition. In this paper, we introduce a robust lane detection method based on a CNN combined with random sample consensus (RANSAC) algorithm. Initially, we calculate edges in an image using a hat shaped kernel, then we detect lanes using the CNN combined with the RANSAC. In the training process of the CNN, input data consists of edge images and target data is images that have real white color lanes on an otherwise black background. The CNN structure consists of 8 layers with 3 convolutional layers, 2 subsampling layers and multi-layer perceptron (MLP) of 3 fully-connected layers. Convolutional and subsampling layers are hierarchically arranged to form a deep structure. Our proposed lane detection algorithm successfully eliminates noise lines and was found to perform better than other formal line detection algorithms such as RANSAC

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.141-151
• /
• 2022

In recent years, the industry and research communities have focused on developing autonomous crack inspection approaches, which mainly include image acquisition and crack detection. In these approaches, mobile devices such as cameras, drones or smartphones are utilized as sensing platforms to acquire structural images, and the deep learning (DL)-based methods are being developed as important crack detection approaches. However, the process of image acquisition and collection is time-consuming, which delays the inspection. Also, the present mobile devices such as smartphones can be not only a sensing platform but also a computing platform that can be embedded with deep neural networks (DNNs) to conduct on-site crack detection. Due to the limited computing resources of mobile devices, the size of the DNNs should be reduced to improve the computational efficiency. In this study, an architecture called pruned crack recognition network (PCR-Net) was developed for the detection of structural cracks. A dataset containing 11000 images was established based on the raw images from bridge inspections. A pruning method was introduced to reduce the size of the base architecture for the optimization of the model size. Comparative studies were conducted with image processing techniques (IPTs) and other DNNs for the evaluation of the performance of the proposed PCR-Net. Furthermore, a modularly designed framework that integrated the PCR-Net was developed to realize a DL-based crack detection application for smartphones. Finally, on-site crack detection experiments were carried out to validate the performance of the developed system of smartphone-based detection of structural cracks.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.11 no.3
• /
• pp.99-112
• /
• 2015

Digital imaging technology has advanced beyond the limits of the multimedia industry IT convergence, and to develop a complex industry, particularly in the field of object recognition, face smart-phones associated with various Application technology are being actively researched. Recently, face recognition technology is evolving into an intelligent object recognition through image recognition technology, detection technology, the detection object recognition through image recognition processing techniques applied technology is applied to the IP camera through the 3D image object recognition technology Face Recognition been actively studied. In this paper, we first look at the essential human factor, technical factors and trends about the technology of the human object recognition based SPFACS(Smile Progress Facial Action Coding System)study measures the smile detection technology recognizes multi-faceted object recognition. Study Method: 1)Human cognitive skills necessary to analyze the 3D object imaging system was designed. 2)3D object recognition, face detection parameter identification and optimal measurement method using the AAM algorithm inside the proposals and 3)Face recognition objects (Face recognition Technology) to apply the result to the recognition of the person's teeth area detecting expression recognition demonstrated by the effect of extracting the feature points.

• Journal of Satellite, Information and Communications
• /
• v.9 no.4
• /
• pp.91-96
• /
• 2014

This paper proposed an adaptive threshold method for detecting flame candidate regions in a infrared image and it adapts according to the contrast and intensity changes in the image. Conventional flame detection systems uses fixed threshold method since surveillance environment does not change, once the system installed. But it needs a adaptive threshold method as requirements of surveillance system has changed. The proposed adaptive threshold algorithm uses the dynamic behavior of flame as featured parameter. The test result is analysed by comparing test result of proposed adaptive threshold algorithm and conventional fixed threshold method. The analysed data shows, the proposed method has 91.42% of correct detection rate and false detection is reduced by 20% comparing to the conventional method.