• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.21 no.6
• /
• pp.30-35
• /
• 2022

This study has focused on the reliability evaluation of FCT inspection automation equipment for electric rice. To evaluate the reliability of FCT inspection automation equipment, voice analysis, Gray/R/G/B channel experiment, FND segment experiment, and robot position repeatability were performed. In the voice analysis experiment, the comparison value between the recorded and digital output waves was over 99%, indicating a very high result. It was confirmed that both the gray/R/G/B experiment using vision and the FND segment could confirm the output value of the product through vision. The position repeatability of the robot is also excellent, so it is concluded that the inspection effect through the FCT automation system will be excellent.

• Journal of information and communication convergence engineering
• /
• v.21 no.2
• /
• pp.139-144
• /
• 2023

Engineers prefer deep neural networks (DNNs) for solving computer vision problems. However, DNNs pose two major problems. First, neural networks require large amounts of well-labeled data for training. Second, the covariate shift problem is common in computer vision problems. Domain adaptation has been proposed to mitigate this problem. Recent work on adversarial-learning-based unsupervised domain adaptation (UDA) has explained transferability and enabled the model to learn robust features. Despite this advantage, current methods do not guarantee the distinguishability of the latent space unless they consider class-aware information of the target domain. Furthermore, source and target examples alone cannot efficiently extract domain-invariant features from the encoded spaces. To alleviate the problems of existing UDA methods, we propose the mixup regularization in adversarial discriminative domain adaptation (ADDA) method. We validated the effectiveness and generality of the proposed method by performing experiments under three adaptation scenarios: MNIST to USPS, SVHN to MNIST, and MNIST to MNIST-M.

• Journal of Drive and Control
• /
• v.21 no.1
• /
• pp.53-58
• /
• 2024

With the recent development of intelligence and automation technologies for construction machinery, the demand for safety and efficiency of road-cutting operations has continued to increase. In response to this, a double-blade road cutter has been developed that can automatically cut roads. However, a double-blade road cutter has a load difference between the two blades due to the ground and wear conditions of the cutting blades. The difference in load between the two blades distorts the direction of travel of the cutter. In this study, a vision sensor-based driving guide technology was developed to correct the driving path of road cutters. In addition, we developed a load-sensing technology that detects blade loads in real-time and controls driving speed in the event of overload.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.6
• /
• pp.1237-1242
• /
• 2017

Computer vision is coming more interesting with new imaging sensors' new capabilities which enable it to understand more its surrounding environment by imitating human vision system with artificial intelligence techniques. In this paper, we made experiments with Kinect camera, a new depth sensor for object detection and distance measurement functions, most essential functions in computer vision such as for unmanned or manned vehicles, robots, drones, etc. Therefore, Kinect camera is used here to estimate the position or the location of objects in its field of view and measure the distance from them to its depth sensor in an accuracy way by checking whether that the detected object is real object or not to reduce processing time ignoring pixels which are not part of real object. Tests showed promising results with such low-cost range sensor, Kinect camera which can be used for object detection and distance measurement which are fundamental functions in computer vision applications for further processing.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.21 no.4
• /
• pp.60-69
• /
• 2022

To keep pace with Industry 4.0, it is imperative for companies to redesign their working environments by adopting robotic automation systems. Automation lines are facilitating the latest cutting-edge technologies, such as 3D vision and industrial robots, to outdo competitors by reducing costs. Considering the nature of the manufacturing industry, a time-saving workflow and smooth linkwork between processes is vital. At Dellics, without any additional new installation in the automation lines, only a few improvements to the working process could raise productivity. Three requirements are the development of gripping technology by utilizing a 3D vision system for the recognition of the material shape and location, research on lighting projectors to target long distances and high illumination, and testing of algorithms/software to improve measurement accuracy and identify products. With some of the functional requisites mentioned above, improved robotic automation systems should provide an improved working environment to maximize overall production efficiency. In this article, the ways in which such a system can become the groundwork for establishing an unmanned working infrastructure are discussed.

• Journal of Industrial Technology
• /
• v.21 no.B
• /
• pp.67-74
• /
• 2001

According as the patterns of PCB (Printed Circuit Board) become denser and complicated, quality and accuracy of PCB influence the performance of final product. It's attempted to obtain trust of 100% about all of parts. Because human inspection in mass-production manufacturing facilities are both time-consuming and very expensive, the automation of visual inspection has been attempted for many years. Thus, automatic visual inspection of PCB is required. In this paper, we used an algorithm which compares the reference PCB patterns and the input PCB patterns are separated an object and a scene by filtering and edge detection. And than compare two image using pattern matching algorithm. We suggest an defect inspection algorithm in PCB pattern, to be satisfied low cost, high speed, high performance and flexibility on the basis of $640{\times}480$ binary pattern.

• Journal of the Semiconductor & Display Technology
• /
• v.10 no.3
• /
• pp.21-28
• /
• 2011

In recent years, owing to the development of the image processing technology, the research to build control system using a vision sensor is stimulated. However, a random time delay must be considered, because it works of a various time to get a result of an image processing in the system. It can be seen as an obstacle factor to a control of visual tracking in real system. In this paper, implementing two vision controllers each, first one is made up PID controller and the second one is consisted of a Smith Predictor, the possibility was shown to overcome a problem of a random time delay in a visual tracking system. A number of simulations and experiments were done to show the validity of this study.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.8 no.2 s.21
• /
• pp.14-21
• /
• 2005

This paper describes the system for detecting the tank crew's(loader's) hand, arm, head and the upper half of the body in a danger area between the turret ceiling and the upper breech mechanism by computer vision-based method. This system informs danger of pressed to death to gunner and commander for the safety of operating mission. The camera mounted ort the top portion of the turret ceiling. The system sets search moving object from this image and detects by using change of image, laplacian operator and clustering algorithm in this area. It alarms the tank crews when it's judged that dangerous situation for operating mission. The result In this experiment shows that the detection rate maintains in $81{\sim}98$ percents.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.6
• /
• pp.21-28
• /
• 2019

Recent years with the development of artificial intelligence and the success of the deep model, they have been deployed in all fields of computer vision. Action recognition, as an important branch of human perception and computer vision system research, has attracted more and more attention. Action recognition is a challenging task due to the special complexity of human movement, the same movement may exist between multiple individuals. The human action exists as a continuous image frame in the video, so action recognition requires more computational power than processing static images. And the simple use of the CNN network cannot achieve the desired results. Recently, the attention model has achieved good results in computer vision and natural language processing. In particular, for video action classification, after adding the attention model, it is more effective to focus on motion features and improve performance. It intuitively explains which part the model attends to when making a particular decision, which is very helpful in real applications. In this paper, we proposed a 3D dense convolutional network based on attention mechanism(ADD-Net), recognition of human motion behavior in the video.

• Journal of the Semiconductor & Display Technology
• /
• v.21 no.1
• /
• pp.75-78
• /
• 2022

Deep Neural Networks (DNN) has become an essential data processing architecture for the implementation of multiple computer vision tasks. Recently, DNN-based algorithms achieve much higher recognition accuracy than traditional algorithms based on shallow learning. However, training and inference DNNs require huge computational capabilities than daily usage purposes of computers. Moreover, with increased size and depth of DNNs, CPUs may be unsatisfactory since they use serial processing by default. GPUs are the solution that come up with greater speed compared to CPUs because of their Parallel Processing/Computation nature. In this paper, we analyze the inference time complexity of DNNs using well-known computer vision library, OpenCV. We measure and analyze inference time complexity for three cases, CPU, GPU-Float32, and GPU-Float16.