• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.4
• /
• pp.1648-1672
• /
• 2020

There are several types of image registration in the sense of stitching separated images that overlap each other. One of these is feature-based registration by a common feature descriptor. In this study, we generate a mosaic of images using feature-based registration for drone aerial images. As a feature descriptor, we apply the scale-invariant feature transform descriptor. In order to investigate the authenticity of the feature points and to have the mapping function, we employ the sample consensus method; we consider the sensed image's inherent characteristic such as the geometric congruence between the feature points of the images to propose a novel hypothesis estimation of the mapping function of the stitching via some optimally chosen initial candidate inliers in the sample consensus method. Based on the experimental results, we show the efficiency of the proposed method compared with benchmark methodologies of random sampling consensus method (RANSAC); the well-ordering property defined in the context and the extensive stitching examples have supported the utility. Moreover, the sample consensus scheme proposed in this study is uncomplicated and robust, and some fatal miss stitching by RANSAC is remarkably reduced in the measure of the pixel difference.

• Journal of Korean Society for Geospatial Information Science
• /
• v.13 no.2 s.32
• /
• pp.65-69
• /
• 2005

Efficient 3D generation of cultural features, such as buildings in urban area is becoming increasingly important for a number of GIS applications. For reconstruction or 3D building in urban area aerial images, satellite images, LIDAR data have been used mainly. In case of automatically extracting and reconstructing of building height using single aerial images or single satellite images, there are a lot of problems, such as mismatching that result from a geometric distortion of optical images. Therefore, researches or integrating optical images and existing 2D GIS data(e.g. digital map) has been in progress. In this paper, we focused on extracting of building height by means or interest points and vortical line locus for reducing matching points. Also we used digital plotter in order to validate for the results in this study using aerial images(1/5,000) and existing digital map(1/1,000).

• Journal of Wetlands Research
• /
• v.16 no.3
• /
• pp.393-401
• /
• 2014

The purpose of this study was to investigate the objective impact in accuracy and reliability with tendency depend on training samples by using the high-resolution images. Supervised classification was performed based on multi-spectral images which made by each satellite and aerial images for considering all of bands' characteristics. The highest accuracy was 84.7% with satellite image(3*3) and 83% with aerial image(5*5) at the accuracy verification phase. Also, the overall accuracy with the consideration of Kappa coefficient were 0.84 for satellite images and 0.82 for aerial images. In all of the images, the smaller training sample was, the higher accuracy showed. Therefore, tree species classification accuracy was tended to rely on training sample size.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.39 no.6
• /
• pp.409-418
• /
• 2021

As the amount of construction for aerial photography increases, the need for automation of quality inspection is emerging. In this study, an experiment was performed to classify or detect clouds in aerial photos using deep learning techniques. Also, classification and detection were performed by including satellite images in the learning data. As algorithms used in the experiment, GoogLeNet, VGG16, Faster R-CNN and YOLOv3 were applied and the results were compared. In addition, considering the practical limitations of securing erroneous images including clouds in aerial images, we also analyzed whether additional learning of satellite images affects classification and detection accuracy in comparison a training dataset that only contains aerial images. As results, the GoogLeNet and YOLOv3 algorithms showed relatively superior accuracy in cloud classification and detection of aerial images, respectively. GoogLeNet showed producer's accuracy of 83.8% for cloud and YOLOv3 showed producer's accuracy of 84.0% for cloud. And, the addition of satellite image learning data showed that it can be applied as an alternative when there is a lack of aerial image data.

• Korean Journal of Soil Science and Fertilizer
• /
• v.49 no.6
• /
• pp.743-750
• /
• 2016

Recently, usage of UAV (Unmanned Aerial Vehicle) has increased in agricultural part. This study was conducted to classify onion and garlic using supervised classification of a fixed-wing UAV (Model : Ebee) images for evaluation of possibility about estimation of onion and garlic cultivation area using UAV images. Aerial images were obtained 11~12 times from study sites in Changryeng-gun and Hapcheon-gun during farming season from 2015 to 2016. The result for accuracy in onion and garlic image classification by R-G-B and R-G-NIR images showed highest Kappa coefficients for the maximum likelihood method. The result for accuracy in onion and garlic classification showed high Kappa coefficients of 0.75~0.97 from DOY 105 to DOY 141, implying that UAV images could be used to estimate onion and garlic cultivation area.

• Korean Journal of Remote Sensing
• /
• v.37 no.4
• /
• pp.747-761
• /
• 2021

With the increasing application of UAVs in intelligent transportation systems, vehicle detection for aerial images has become an essential engineering technology and has academic research significance. In this paper, a vehicle detection method for aerial images based on the YOLOv4 deep learning algorithm is presented. At present, the most known datasets are VOC (The PASCAL Visual Object Classes Challenge), ImageNet, and COCO (Microsoft Common Objects in Context), which comply with the vehicle detection from UAV. An integrated dataset not only reflects its quantity and photo quality but also its diversity which affects the detection accuracy. The method integrates three public aerial image datasets VAID, UAVD, DOTA suitable for YOLOv4. The training model presents good test results especially for small objects, rotating objects, as well as compact and dense objects, and meets the real-time detection requirements. For future work, we will integrate one more aerial image dataset acquired by our lab to increase the number and diversity of training samples, at the same time, while meeting the real-time requirements.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.1917-1920
• /
• 2003

Automatic tracking of a moving object such as a person is a demanding technique especially in surveillance. This paper describes an experimental system for tracking a moving object on the ground by using a visually controlled aerial robot. A blimp is used as the aerial robot in the proposed system because of its locality in motion and its silent nature. The developed blimp is equipped with a camera for taking downward images and four rotors for controlling the progression. Once a camera takes an image of a specified moving object on the ground, the blimp is controlled so that it follows the object by the employment of the visual information. Experimental results show satisfactory performance of the system. Advantages of the present system include that images from the air often enable us to avoid occlusion among objects on the ground and that blimp’s progression is much less restricted in the air than, e.g., a mobile robot running on the ground.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.2
• /
• pp.146-156
• /
• 1998

Thispapr proposes hybrid navigation parameter estimation using sequential aerial images. The proposed navigation parameter estimation system is composed of two parts: relative position estimation and absolute position estimation. the relative position estimation recursively computes the current velocity and absolute position estimation. The relative position estimation recursively computes the current velocity and position of an aircraft by accumulating navigation parameters extracted from two succesive aerial images. Simple accumulation of parameter values decreases reliability of the extracted parameters as an aircraft goes on navigating. therefore absolute position estimation is required to compensate for position error generated in the relative position step. The absolute position estimation algorithm combining image matching and digital elevation model(DEM) matching is presented. Computer simulation with real aerial image sequences shows the efficiency of the proposed hybrial algorithm.

• Proceedings of the IEEK Conference
• /
• 2009.05a
• /
• pp.216-218
• /
• 2009

This paper proposes an image processing algorithm to stabilize shaken scenes of UAV(Unmanned Aerial Vehicle) caused by vehicle self-vibration and aerodynamic disturbance. The proposed method stabilizes images by compensating estimated shake motion which is evaluated from global motion. The global motion between two continuous images modeled by 6 parameter warping model is estimated by non-linear square method based on Gauss-Newton algorithm with excluding outlier region. The shake motion is evaluated by subtracting the global motion from aerial vehicle motion obtained by averaging global motion. Experimental results show that the proposed method stabilize shaken scenes effectively.

• Korean Journal of Remote Sensing
• /
• v.34 no.6_1
• /
• pp.893-903
• /
• 2018

The necessity of automatic precise georeferencing is increasing with the increase applications of high-resolution satellite imagery. One of the methods for collecting ground control points (GCPs) for precise georeferencing is to use chip images obtained by extracting a subset of an image map such as an ortho-aerial image, and can be automated using an image matching technique. In this case, the importance of the image matching success rate is increased due to the limitation of the number of the chip images for the known reference points such as the unified control point. This study aims to propose a method to improve the success rate of image matching between KOMPSAT-3A images and GCP chip images from aerial ortho-images. We performed the image matching with 7 cases of band pair using KOMPSAT-3A panchromatic (PAN), multispectral (MS), pansharpened (PS) imagery and GCP chip images, then compared matching success rates. As a result, about 10-30% of success rate is increased to about 40-50% when using PS imagery by using PAN and MS imagery. Therefore, using PS imagery for image matching of KOMPSAT-3A images and aerial ortho-images would be helpful to improve the matching success rate.