• The Korean Journal of Applied Statistics
• /
• v.23 no.2
• /
• pp.305-315
• /
• 2010

In this paper, a frequentistic approach to impute the values of R, G, B-components in random missing pixels of color image is provided. Under assumption that the given image is a realization of Gaussian Markov random field, its model is designed such that each neighbor pixel values for a given pixel follows (independently) the normal distribution with covariance matrix scaled by an evaluates of the similarity between two pixel values, so that the imputation is not to be affected by the neighbors with different color. An approximate EM-based algorithm maximizing the underlying likelihood is implemented to estimate the parameters and to impute the missing pixel values. Some experiments are presented to show its effectiveness through performance comparison with a popular interpolation method.

• 전자공학회논문지 IE
• /
• v.44 no.2
• /
• pp.32-40
• /
• 2007

In this paper, the crease extraction method is proposed to improve the accuracy of feature extraction within the fingerprint image. First of all, for each pixel in fingerprint image, it calculates the average grey level and variance to determine if the current pixel composes the crease, and estimates the direction of crease. Secondly, once the direction of every pixel in crease candidate area is estimated, it is decomposed into 8 different images, depending on their direction. The properties of crease consists of the length of the crease candidate area, the correspondence between the crease direction and the pixel distribution direction, the difference between the ridge direction and the pixel distribution direction, and finally the grey level of the candidate pixels. The proposed method finally extracts the crease from the crease clusters estimated from directional images. In conclusion, applying the proposed method improved the accuracy of overall feature extraction by 91.4% by accurately and precisely extracting the crease from fingerprint image.

• Korean Journal of Remote Sensing
• /
• v.38 no.1
• /
• pp.33-43
• /
• 2022

This study presents a method to restore an optical satellite image with distortion and occlusion due to fog, haze, and clouds to one that minimizes degradation factors by referring to the same type of peripheral image. Specifically, the time and cost of re-photographing were reduced by partially occluding a region. To maintain the original image's pixel value as much as possible and to maintain restored and unrestored area continuity, a simulation restoration technique modified with the Cycle Generative Adversarial Network (CycleGAN) method was developed. The accuracy of the simulated image was analyzed by comparing CycleGAN and histogram matching, as well as the pixel value distribution, with the original image. The results show that for Site 1 (out of three sites), the root mean square error and R² of CycleGAN were 169.36 and 0.9917, respectively, showing lower errors than those for histogram matching (170.43 and 0.9896, respectively). Further, comparison of the mean and standard deviation values of images simulated by CycleGAN and histogram matching with the ground truth pixel values confirmed the CycleGAN methodology as being closer to the ground truth value. Even for the histogram distribution of the simulated images, CycleGAN was closer to the ground truth than histogram matching.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.231-233
• /
• 2003

Agricultural land use generally shows specific temporal characteristics of NDVI obtained from satellite data. In terms of winter wheat, a higher value compared with other land use types in May and a considerably low value in June could be discriminative features of temporal change of NDVI. In this study, the author examined methods for estimating winter wheat sown area in sub-pixel level of coarse resolution satellite data using temporal characteristics of NDVI. Application of the methods to the major grain production area in China exhibited properly a spatial distribution pattern of winter wheat sown area.

• The Bulletin of The Korean Astronomical Society
• /
• v.41 no.2
• /
• pp.35.1-35.1
• /
• 2016

To understand the coevolution of Brightest Cluster Galaxies (BCGs) and their host clusters, we conduct a case study on the BCGs in dynamically young and old clusters, Abell 1139 (A1139) and Abell 2589 (A2589). We analyze the pixel color-magnitude diagrams (pCMDs) using deep g- and r-band images, obtained from the CFHT observations. (1) While the overall shapes of the pCMDs are similar to those of typical early-type galaxies, the A2589-BCG tends to have redder mean pixel color and smaller pixel color deviation at given surface brightness than the A1139-BCG. (2) The mean pixel color distribution as a function of pixel surface brightness indicates that the A2589-BCG formed a larger central body by major dry mergers at an early epoch than the A1139-BCG, while they have grown commonly by subsequent minor mergers. (3) The spatial distributions of the pixels with deviated colors reveal that the A1139-BCG experienced considerable tidal events more recently than the A2589-BCG, whereas the A2589-BCG has an asymmetric compact core possibly resulting from major dry merger at an early epoch. (4) The A2589-BCG shows a very large faint-to-bright pixel number ratio compared to early-type non-BCGs, whereas the ratio for the A1139-BCG is not distinctively large. These results imply that the BCG in the dynamically older cluster (A2589) formed earlier and is relaxed better.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.12 no.3
• /
• pp.29-35
• /
• 2012

In this paper, it is presented the method of image recovering which existing is only pixel processing, but suggesting method is concluding image clustering overlap degree after classfying around unit fixel to crowd pixel. Concluding overlap degree threshold value is after identifying pattern pixel and grasping geometry structure of sample pattern and deduction of deciding function. distinguishing feature space is above four dimension is reason of not visual observation of pattern structure. consideration of distribution structure is distance of center of crowd pixel, the number of each crowd pattern pixel and standard deviation. The over threshold value elimate the overlap image and the downward is recovered and established dynamic image. memory storage deduction of 20% and elevation of 15% performance is estimated in recovery of image.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.16 no.1
• /
• pp.75-84
• /
• 1998

To achieve more accurate information from satellite image data, a research on a technique for mixed pixel ex-traction has been produced. The mixed pixels with only two land covers have been experimented. By analyzing canonical vector in canonical correlation classification, the mixed pixels have been classified. The ratio of the two canonical weighted values-the elements of canonical vector have been used as a threshold to discriminate mixed pixels. In case of the classification for the mixed pixels of bridge and water class in TM data before or after the 1st of September, the threshold for the optimal classification of the mixed pixels is 4.0. That is, if the ratio of the two canonical weighted values is less than 4.0, the pixel is a mixed pixel. Also, using the distribution of canonical weighted values, the constitution percentages of land covers within one mixed pixel can be approximately deducted. The accuracy of mixed pixel extraction for experimental area is 90％ and quite acceptable. Conclusively, a technique for mixed pixel extraction by canonical vector analysis is effective.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1526-1526
• /
• 2001

To improve the accuracy of sweetness sensor in automated sorting operations, it is necessary to clarify unevenness of the sugar content distribution within fruits. And it is expected that the technique to evaluate the content distribution in fruits contribute to the development of the near-infrared (NIR) imaging spectroscopy. Sugiyama (1999) had succeeded to visualize the distribution of the sugar content on the surface of a half-cut green fresh melon. However, this method cannot be applied to red flesh melons because it depends on information of the absorption band of chlorophyll (676 nm), which is affected by the color of the fresh. The objective of this study was to develop the universal visualization method depends on the absorption band of sugar, which can be applied to various kinds of melons and other fruits. The relationship between the sugar contents and absorption spectra of both green and red fresh melons were investigated by using a NIR spectrometer to determine the absorption band of sugar. The combination of 2$\^$nd/ derivative absorbances at 902 nm and 874 nm was highly correlated with the sugar contents. The wavelength of 902 nm is attributed to the absorption band of sugar. A cooled charge-coupled device (CCD) imaging camera which has 16 bit (65536 steps) A/D resolution was equipped with rotating band-pass filter wheel and used to capture the spectral absorption images of the flesh of a vertically half-cut red fresh melon. The advantage of the high A/D resolution in this research is that each pixel of the CCD is expected to function as a detector of the NIR spectrometer for quantitative analysis. Images at 846 nm, 874 nm, 902 nm and 930 nm were acquired using this CCD camera. Then the 2$\^$nd/ derivative absorbances at 902 nm and 874 nm at each pixel were calculated using these four images. On the other hand, parts of the same melon were extracted for capturing the images and squeezed for the measurement of sugar content. Then the calibration curve between the combination of 2$\^$nd/ derivative absorbances at 902 nm and 874 nm and sugar content was developed. The calibration method based on NIR spectroscopy techniques was applied to each pixel of the images to convert the 2$\^$nd/ derivative absorbances into the Brix sugar content. Mapping the sugar content value of each pixel with linear color scale, the distribution of the sugar content was visualized. As a result of the visualization, it was quantitatively confirmed that the Brix sugar contents are low at the near of the skin and become higher towards the seeds. This result suggests that the visualization technique by the NIR imaging spectroscopy could become a new useful method fer quality evaluation of melons.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.32A no.12
• /
• pp.122-129
• /
• 1995

The most important feature of a-Si TFT is dense localized states such as dangling bonds which exist in tis bandgap. Electrons trapped by localized states dominate the potential distribution in the active a-Si region ,and influence the performance of a-Si TFT. In this paper, we describe the electrical characteristics of a-Si TFT with respect to trap distribution within bandgap, electron mobility and interface states using 2-Dimensional device simulator and compare the result of simulation with measurements. Using the mixed-mode simulator, we can predict the potential variation of pixel which causes residual image problem during the turn-off of a-Si TFT driving circuit. Therefore it is possible to consider trade-off between potential variation of pixel and turn-on current of a-Si TFT for the optimized driving circuit.

• The Bulletin of The Korean Astronomical Society
• /
• v.36 no.2
• /
• pp.103.2-103.2
• /
• 2011

We measure the degree of polarization of the lunar regolith to map the distributions of the age and the particle size. We use a 12cm refracting telescope with a 2k-square pixel color CCD (R band) and a polarization filter. The angular resolution obtained is 3.02 km/pixel. Our goal is to obtain a map of the lunar particle size distribution on the lunar regolith and then that of the age distribution. Polarization of the light scattered by lunar surface contains information on their mean particle size. The mean particle size of the lunar surface has been decreased by continued micro-meteoroid impact over a long period. One can estimate the age of the lunar surface if the mean particle size is known. Particle sizes can be measured through observations of polarization because the mean particle size is related to the maximum polarization and albedo. The age and the particle size of the lunar regolith can give vital information for the future lunar exploration.