• 대한전기학회논문지:시스템및제어부문D
• /
• 제52권4호
• /
• pp.212-222
• /
• 2003

The ability to efficiently and robustly recover accurate 3D terrain models from sets of stereoscopic images is important to many civilian and military applications. A stereo matching has been an important tool for reconstructing three dimensional terrain. However, there exist many factors causing stereo matching error, such as occlusion, no feature or repetitive pattern in the correlation window, intensity variation, etc. Among them, occlusion can be only resolved by true multi-image stereo. In this paper, we present multi-image stereo method using DEM fusion as one of efficient and reliable true multi-image methods. Elevations generated by all pairs of images are combined by the fusion process which accepts an accurate elevation and rejects an outlier. We propose three fusion schemes: THD(Thresholding), BPS(Best Pair Selection) and MS(Median Selection). THD averages elevations after rejecting outliers by thresholding, while BPS selects the most reliable elevation. To determine the reliability of a elevation or detect the outlier, we employ the measure of self-consistency. The last scheme, MS, selects the median value of elevations. We test the effectiveness of the proposed methods with a quantitative analysis using simulated images. Experimental results indicate that all three fusion schemes showed much better improvement over the conventional binocular stereo in natural terrain of 29 Palms and urban site of Avenches.

• 정보처리학회논문지B
• /
• 제17B권2호
• /
• pp.115-124
• /
• 2010

최근 스테레오 정합 기술은 정합하고자 하는 픽셀을 포함한 국부적인(local) 영상의 정합 비용과 시차의 변화 비용을 누적하는 전역적(global)인 방법을 많이 사용하고 있다. 특히 전역적 스테레오 정합에서도 비용누적 (cost accumulation)의 방향을 일반적인 수평방향이 아닌 다수의 방향을 사용하는 연구가 늘고 있다. 본 논문에서는 기존의 스테레오 정합 기술을 다중 방향성 정합 기술로 확장하는 방법을 제안한다. 픽셀의 국부적인 정합 비용은 단순한 NCC (Normalized Cross Correlation)를 사용하였고 전역적 정합 기술의 하나인 정합선 최적화(Scan-line Optimization) 방법을 다중 방향으로 확장하는 기술을 제안하였다. 우선 정합선 최적화를 다중 방향으로 실행한 후 이들 결과를 이용하여 신뢰도가 높은 시차영상 (disparity image)을 획득한다. 반복적인 다중 방향 정합선 최적화 시행 후, 시차영상에서 남은 공백은 홀 복원 방법으로 계산한다. 시차가 구해진 픽셀에 대해서는 신뢰도 점수를 매긴 다음 이 점수를 확산하여 신뢰도 점수 테이블에서 가장 높은 값을 가지는 시차값으로 홀을 복원하였다. 제안하는 기술을 미들버리(Middlebury)의 스테레오 영상을 사용하여 오차를 분석하였다. 기존의 전역적 방법과 제안 기술을 이용하여 시차영상을 계산하고 그 오차를 비교하였다.

• Journal of Mechanical Science and Technology
• /
• 제17권10호
• /
• pp.1411-1422
• /
• 2003

The detection of free spaces between obstacles in a scene is a prerequisite for navigation of a mobile robot. Especially for stereo vision-based navigation, the problem of correspondence between two images is well known to be of crucial importance. This paper describes multi-range approach of area-based stereo matching for grid mapping and visual navigation in uncertain environment. Camera calibration parameters are optimized by evolutionary algorithm for successful stereo matching. To obtain reliable disparity information from both images, stereo images are to be decomposed into three pairs of images with different resolution based on measurement of disparities. The advantage of multi-range approach is that we can get more reliable disparity in each defined range because disparities from high resolution image are used for farther object a while disparities from low resolution images are used for close objects. The reliable disparity map is combined through post-processing for rejecting incorrect disparity information from each disparity map. The real distance from a disparity image is converted into an occupancy grid representation of a mobile robot. We have investigated the possibility of multi-range approach for the detection of obstacles and visual mapping through various experiments.

• Korean Journal of Geomatics
• /
• 제1권1호
• /
• pp.43-49
• /
• 2001

Although many kinds of stereo matching method have been developed in the field of computer vision and photogrammetry, wrong matches are not easy to avoid. This paper presents a new method to reduce wrong matches after matching, and experimental results are reported. The main idea is to analyze the histogram of the image attribute differences between each pair of image patches matched. Typical image attributes of image patch are the mean and the standard deviation of gray value for each image patch, but there could be other kinds of image attributes. Another idea is to check relative position among potential matches. This paper proposes to use Gaussian blunder filter to detect the suspicious pair of candidate match in relative position among neighboring candidate matches. If the suspicious candidate matches in image attribute difference or relative position are suppressed, then many wrong matches are removed, but minimizing the suppression of good matches. The proposed method is easy to implement, and also has potential to be applied as post processing after image matching for many kinds of matching methods such as area based matching, feature matching, relaxation matching, dynamic programming, and multi-channel image matching. Results show that the proposed method produces fewer wrong matches than before.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 Ⅳ
• /
• pp.2295-2298
• /
• 2003

The correspondence problem for stereo image matching plays an important role in expanding view points as multi view video applications become more popular. The conventional disparity estimation algorithms have limitation to find exact disparities because they consider not image features but similiar intensity points. Thus we propose an efficient disparity estimation algorithm considering features of stereo image pairs. As simulation results, our proposed method confirms better intermediate views than the existing block-matching methods.

• 융합신호처리학회논문지
• /
• 제13권1호
• /
• pp.12-18
• /
• 2012

스테레오 대응성은 스테레오 비전에서 중요한 문제이다. 본 논문은 다해상도 기법과 AD-Census를 이용한 유전 알고리즘 기반의 스테레오 정합 기법을 제안한다. 정합 환경을 최적화 문제로 간주하여 유전 알고리즘으로 변위를 탐색한다. 그리고 에지 픽셀을 이용한 적응적 염색체 구조와 교배 방식을 적용한다. 비용함수는 스테레오 정합에서 주로 고려할 수 있는 제약 조건으로 구성하였고, 변위오차를 줄이기 위해 AD-Census 척도를 사용하였다. 처리의 효율을 높이기 위해 영상 피라미드 방법을 적용하여 최저해상도에서 최초 변위 도를 계산한다. 그리고 최초 변위도는 다음 해상도 단계로 전파되어, 보간된 후 지역 특징 벡터를 이용하여 정제를 수행한다. 실험을 통해 제안한 방법이 다른 유전 알고리즘 기반 기법들에 비해 변위 탐색 시간을 감소시킬 뿐만 아니라 정합의 타당성을 보증함을 확인하고자 한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권6호
• /
• pp.3208-3229
• /
• 2017

Rectification is an essential procedure for simplifying the disparity extraction of stereo matching algorithms by removing vertical mismatches between left and right images. To support real-time stereo matching, studies have introduced several look-up table (LUT)- and computational logic (CL)-based rectification approaches. However, to support high-resolution images, the LUT-based approach requires considerable memory resources, and the CL-based approach requires numerous hardware resources for its circuit implementation. Thus, this paper proposes a multi-level accumulation-based rectification method as a simple CL-based method and its circuit implementation. The proposed method, which includes distortion correction, reduces addition operations by 29%, and removes multiplication operations by replacing the complex matrix computations and high-degree polynomial calculations of the conventional rectification with simple multi-level accumulations. The proposed rectification circuit can rectify $1,280{\times}720$ stereo images at a frame rate of 135 fps at a clock frequency of 125 MHz. Because the circuit is fully pipelined, it continuously generates a pair of left and right rectified pixels every cycle after 13-cycle latency plus initial image buffering time. Experimental results show that the proposed method requires significantly fewer hardware resources than the conventional method while the differences between the results of the proposed and conventional full rectifications are negligible.

• 대한임베디드공학회논문지
• /
• 제16권5호
• /
• pp.179-186
• /
• 2021

For stereo matching based on deep learning, the design of network structure is crucial to the calculation of matching cost, and the time-consuming problem of convolutional neural network in image processing also needs to be solved urgently. In this paper, a method of stereo matching using sparse loss volume in parallax dimension is proposed. A sparse 3D loss volume is constructed by using a wide step length translation of the right view feature map, which reduces the video memory and computing resources required by the 3D convolution module by several times. In order to improve the accuracy of the algorithm, the nonlinear up-sampling of the matching loss in the parallax dimension is carried out by using the method of multi-category output, and the training model is combined with two kinds of loss functions. Compared with the benchmark algorithm, the proposed algorithm not only improves the accuracy but also shortens the running time by about 30%.

• 제어로봇시스템학회논문지
• /
• 제9권11호
• /
• pp.899-905
• /
• 2003

An intermediate images generation method using multi-resolution based hierarchical block matching disparity map is proposed. This method is composed of a disparity estimation, an occlusion detection and intermediate image synthesis. For the disparity estimation, which is one of the important processes in intermediate image synthesis, we use the multi-resolution based hierarchical block matching algorithm to overcome the imperfect ness of block matching algorithm. The proposed method makes disparity maps more accurate and dense by multi-resolution based hierarchical block matching, and the estimated disparity maps are used to generate intermediate images of stereo images. Generated intermediate images show 0.1∼1.4 ㏈ higher PSNR than the images obtained by block matching algorithm.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제53권4호
• /
• pp.277-285
• /
• 2004

We present a new hybrid stereo matching technique in terms of the co-operation of area-based stereo and feature-based stereo. The core of our technique is that feature matching is carried out by the reference of the disparity evaluated by area-based stereo. Since the reference of the disparity can significantly reduce the number of feature matching combinations, feature matching error can be drastically minimized. One requirement of the disparity to be referenced is that it should be reliable to be used in feature matching. To measure the reliability of the disparity, in this paper, we employ the self-consistency of the disunity Our suggested technique is applied to the detection of 3D line segments by 2D line matching using our hybrid stereo matching, which can be efficiently utilized in the generation of the rooftop model from urban imagery. We carry out the experiments on our hybrid stereo matching scheme. We generate synthetic images by photo-realistic simulation on Avenches data set of Ascona aerial images. Experimental results indicate that the extracted 3D line segments have an average error of 0.5m and verify our proposed scheme. In order to apply our method to the generation of 3D model in urban imagery, we carry out Preliminary experiments for rooftop generation. Since occlusions are occurred around the outlines of buildings, we experimentally suggested multi-image hybrid stereo system, based on the fusion of 3D line segments. In terms of the simple domain-specific 3D grouping scheme, we notice that an accurate 3D rooftop model can be generated. In this context, we expect that an extended 3D grouping scheme using our hybrid technique can be efficiently applied to the construction of 3D models with more general types of building rooftops.