• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.901-905
• /
• 2005

Monte Carlo localization is known to be one of the most reliable methods for pose estimation of a mobile robot. Much research has been done to improve performance of MCL so far. Although MCL is capable of estimating the robot pose even for a completely unknown initial pose in the known environment, it takes considerable time to give an initial estimate because the number of random samples is usually very large especially for a large-scale environment. For practical implementation of the MCL, therefore, a reduction in sample size is desirable. This paper presents a novel approach to reducing the number of samples used in the particle filter for efficient implementation of MCL. To this end, the topological information generated off- line using a thinning method, which is commonly used in image processing, is employed. The topological map is first created from the given grid map for the environment. The robot scans the local environment using a laser rangefinder and generates a local topological map. The robot then navigates only on this local topological edge, which is likely to be the same as the one obtained off- line from the given grid map. Random samples are drawn near the off-line topological edge instead of being taken with uniform distribution, since the robot traverses along the edge. In this way, the sample size required for MCL can be drastically reduced, thus leading to reduced initial operation time. Experimental results using the proposed method show that the number of samples can be reduced considerably, and the time required for robot pose estimation can also be substantially decreased.

• Journal of IKEEE
• /
• v.14 no.3
• /
• pp.224-231
• /
• 2010

Automatic detection of objects in images has been one of core challenges in the areas such as computer vision and pattern analysis. Especially, with the recent deployment of personal mobile devices such as smart phone, such technology is required to be transported to them. Usually, these smart phone users are equipped with devices such as camera, GPS, and gyroscope and provide various services through user-friendly interface. However, the smart phones fail to give excellent performance due to limited system resources. In this paper, we propose a new scheme to improve object recognition performance based on pre-computation and simple local features. In the pre-processing, we first find several representative parts from similar type objects and classify them. In addition, we extract features from each classified part and train them using regression functions. For a given query image, we first find candidate representative parts and compare them with trained information to recognize objects. Through experiments, we have shown that our proposed scheme can achieve resonable performance.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.8C
• /
• pp.583-590
• /
• 2008

In this paper, we present a spatially adaptive post-processing algorithm for H.264 video coding standard to remove blocking artifacts. The loop filter of H.264 increases computational complexity of the encoder. Furthermore it doesn't clearly remove the blocking artifacts, resulting in over-blurring. For overcoming them, we combine the projection method with the Constraint Least Squares(CLS) method to restore the high quality image. To reflect the Human Visual System, we adopt the weight norm CLS method. Particularly pixel location-based local variance and laplacian operator are newly defined for the CLS method. In addition, the fact that correlation among adjoining pixels is high is utilized to constrain the solution space when the projection method is applied. Quantization Index(QP) of H.264 is also used to control the degree of smoothness. The simulation results show that the proposed post-processing filter works better than the loop filter of H.264 and converges more quickly than the CLS method.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.5
• /
• pp.186-201
• /
• 2021

As autonomous driving technology advances, the accuracy of the vehicle position is important for recognizing the environments around driving. Map-matching localization techniques based on high definition (HD) maps have been studied to improve localization accuracy. Because conventional map-matching techniques estimate the vehicle position based on an HD map reference dataset representing the center of the lane, the estimated position does not reflect the deviation of the lateral distance within the lane. Therefore, this paper proposes a localization system based on the reference lateral position dataset extracted using image processing and HD maps. Image processing extracts the driving lane number using inverse perspective mapping, multi-lane detection, and yellow central lane detection. The lane departure method estimates the lateral distance within the lane. To collect the lateral position reference dataset, this approach involves two processes: (i) the link and lane node is extracted based on the lane number obtained from image processing and position from GNSS/INS, and (ii) the lateral position is matched with the extracted link and lane node. Finally, the vehicle position is estimated by matching the GNSS/INS local trajectory and the reference lateral position dataset. The performance of the proposed method was evaluated by experiments carried out on a highway environment. It was confirmed that the proposed method improves accuracy by about 1.0m compared to GNSS / INS, and improves accuracy by about 0.04m~0.21m (7~30%) for each section when compared with the existing lane-level map matching method.

• The KIPS Transactions:PartB
• /
• v.17B no.5
• /
• pp.363-374
• /
• 2010

In this paper, we propose the method for extracting Homogeneity Threshold($H_T$) and for segmenting homogeneous regions by USRG(Unseeded Region Growing) with $H_T$. The $H_T$ is a criterion to distinguish homogeneity in neighbor pixels and is computed automatically from the original image by proposed method. Theoretical background for proposed method is based on the Otsu's single level threshold method. The method is used to divide a small local part of original image int o two classes and the sum($\sigma_c$) of standard deviations for the classes to satisfy special conditions for distinguishing as different regions from each other is used to compute $H_T$. To find validity for proposed method, we compare the original image with the image that is regenerated with only the segmented homogeneous regions and show up the fact that the difference between two images is not exist visually and also present the steps to regenerate the image in order the size of segmented homogeneous regions and in order the intensity that includes pixels. Also, we show up the validity of proposed method with various results that is segmented using the homogeneity thresholds($H^*_T$) that is added a coefficient ${\alpha}$ for adjusting scope of $H_T$. We expect that the proposed method can be applied in various fields such as visualization and animation of natural image, anatomy and biology and so on.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.2
• /
• pp.155-160
• /
• 2014

There are many image processing based algorithms and applications that fail when illumination change occurs. Therefore, the illumination change has to be detected then the illumination change occurred images need to be enhanced in order to keep the appropriate algorithm processing in a reality. In this paper, a new method for detecting illumination changes efficiently in a real time by using local region information and fuzzy logic is introduced. The effective way for detecting illumination changes in lighting area and the edge of the area was selected to analyze the mean and variance of the histogram of each area and to reflect the changing trends on previous frame's mean and variance for each area of the histogram. The ways are used as an input. The changes of mean and variance make different patterns w hen illumination change occurs. Fuzzy rules were defined based on the patterns of the input for detecting illumination changes. Proposed method was tested with different dataset through the evaluation metrics; in particular, the specificity, recall and precision showed high rates. An automatic parameter selection method was proposed for contrast limited adaptive histogram equalization method by using entropy of image through adaptive neural fuzzy inference system. The results showed that the contrast of images could be enhanced. The proposed algorithm is robust to detect global illumination change, and it is also computationally efficient in real applications.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.7
• /
• pp.1383-1388
• /
• 2016

Although image signal processing is used in many fields, degradation takes place in the process of transmitting image data by several causes. CWMF, A-TMF, and AWMF are the typical methods to eliminate noises from image data damaged under salt and pepper noise environment. However, those filters are not effective for noise rejection under highly dense noise environment. In this respect, the present study proposed an algorithm to remove in salt and pepper noise. In case the center pixel is determined to be non-noise, it is replaced with original pixel. In case the center pixel is noise, it segments local mask into 4 directions and uses linear interpolation to estimate original pixel. And then it applies spatial weight to the estimated pixel. The proposed algorithm shows a high PSNR of 24.56[dB] for House images that had been damaged of salt and pepper noise(P = 50%), compared to the existing CWMF, A-TMF and AWMF there were improvements by 16.46[dB], 12.28[dB], and 12.32[dB], respectively.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.5
• /
• pp.1123-1131
• /
• 1996

Several improved quadrature polar separable (QPS) filters have been proposed and applied in texture processing since Knutsson proposed the QPS filter. They include a Knutsson's cosine function or oan exponential attenuation function, as the orientational function, and a Knutsson's exponential function or a finite prolate spheroidal sequence (FPSS) or an asymptotic FPSS, as the radial weighting functions. They represent different properties in terms of the generation of texture images, the orientational estimation, and the segmentation of synthetic texture image. In this paper, we have constructed several kernal functions for the 2-D QPS filter and analyzed their properties. A series of experiments have been carried out in order to estimate the frequency components and orientational angles of a local texture in Fourier domain. finally some problems encountered in applying QPS filters to feature description and segmentation are considered. Experimental results show that the improved Knutsson's filter and the asymptotic FPSS filter are useful in terms of the orientational estimation and the sementation of synthetic texture image.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.10
• /
• pp.1284-1289
• /
• 2016

Recently, digital image processing is applied a lot to the broadcasting, communication, computer graphic, and medical sectors. It generates noise when data is transmitted. There are many kinds of noises that add to the image such as salt and pepper noise, AWGN, and complex noise. Thus, this study divides the corrupted image into four4 areas and estimates the types of noises each pixel, and this study suggested a switching filter that separates the estimated into salt and pepper noise and AWGN. In the case that center pixel of local mask is corrupted by salt and pepper noise, it used a histogram probability weighting of subdivided area. Also, in case that it is corrupted by AWGN, algorithm that is applied to with different weights given for the distribution of each area with using subdivided area's distribution was suggested. For an objective comparison and conclusion, this study used PSNR and compared to existing methods.

• Korean Journal of Remote Sensing
• /
• v.26 no.5
• /
• pp.489-495
• /
• 2010

Hyperspectral imaging obtains information with a wider wavelength range a large number of bands. However, a high correlation between each band, computation cost, and noise causes inaccurate results in cases of no pre-processing. The noises of band extraction and elimination positively necessary in hyperspectral imaging. Since the previous studies have used a characteristic the whole image, a local characteristic of the image is considered for the noise band extraction. In this study, the Quadtree, which is a data structure algorithm. and the fractal dimension are adopted for noise band extraction in Hyperion images. The fractal dimensions of the segments divided by the Quadtree structure are calculated, and variation is used. We focused on the extraction of random noise bands in Hyperion images and compared them with the reference data made by visual decisions. The proposed algorithm extracts the most bands, including random noises. It is possible to eliminate more than 30 noise bands, regardless of images.