http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
이건일,김덕규 경북대학교 전자기술연구소 1981 電子技術硏究誌 Vol.2 No.1
Microprocessor-controlled digital speed control system of chopper-fed do series motor has been designed. By keeping speed constant, the actual speed is measured and compared in the microprocessor with the desired speed setting which was given from external toggle switch or keyboard. Then produced error is used to adjust the 'ON' time of SCR chopper and the output voltage of the programmable power supply such that motor speed has fast response and is maintained at its desired value. The prototype's transient response is improved than the system without control and steady-state error is less than ±25%.
에지영역을 보상한 영역분류 대역간 예측 벡터 양자화를 이용한 다분광 화상데이타의 압축
김영춘,최승태,정원식,이건일,김경규 경북대학교 전자기술연구소 1997 電子技術硏究誌 Vol.18 No.2
In this paper, we proposed a multispectral image data compression method using classified interband predictive VQ with edge region compensation. The proposed method classifies each pixel vector. For each class, we perform classified intraband VQ and classified interband predictive VQ. In edge region case, edge region compensates using class information and gray value of neighboring blocks of quantized reference band. Then we perform two methods which are classified interband predictive VQ with edge region compensation and without edge region compensation. And we choose what has smaller reconstruction error in results of two methods. As a result of the proposed method, we can achieve better compression efficiency than conventional method.
선택적인 대역간 예측 및 VQ를 이용한 인공위성 화상데이타 압축
김영춘,이건일,김경규,이승진 경북대학교 전자기술연구소 1996 電子技術硏究誌 Vol.17 No.2
In this paper, we propose a satellite image data compression method using selective interband prediction/vector quantization(VQ) based on classification. This method classifies each pixel vector considering spectral characteristics of multispectral images. Then we perform the classified VQ to remove spatial redundancy for a reference band image that has the highest correlation spatially. To remove spectral redundancy, we perform the classified interband prediction for the band images that have high correlation spectrally and perform the classified interband for the remaining band images. Iy selecting the interband prediction and VQ for spectral images effectively, this method removes effectively the spectral redundancy in multispectral images and improves compression efficiency. Experiments on Landsat The multispectral image show that compression efficiency of the proposed method is better than that of the conventional Gupta's method.
블럭 분류를 이용한 웨이브릿 변환 영역에서의 선별적인 움직임 추정
김영춘,정원식,이건일,권성근,이승진 경북대학교 전자기술연구소 1997 電子技術硏究誌 Vol.18 No.2
Wavelet transform has recently emerged as a promising technique for video processing application due to its flexibility in representing nonstationary video signals. In this paper, we propose a selective motion estimation method using block classification in wavelet transform domain. The proposed method contrasts with the multiresolution motion estimation (MRME) scheme where motion estimation is done for entire image block. The proposed method classifies each block in the baseband into one of the four classes, non-edge, horizontal-edge, vertical-edge, and mixed-edge block, using wavelet coefficients of the horizontal and vertical subbands in the lowest resolution layer. According to these block classes, the blocks in the higher frequency layer are selectively motion-estimated. Because motion estimation is selectively done, this method reduces cost of calculation in block matching and bit rate in motion vector transmission. Computer simulation show that the proposed scheme has considerable reduction in the bit rate which results in significant improvements in coding performance of the proposed wavelet coding compared to the MRME based wavelet coding for video compression.
프레임간 예측 및 인접블럭의 움직임 벡터를 이용한 예측 움직임 추정기법
김영춘,정원식,이건일,김경규,강규창,이승진 경북대학교 전자기술연구소 1997 電子技術硏究誌 Vol.18 No.1
In this paper, we propose predictive motion estimation algorithm which can predict motion without additional side information considering spatio-temporal correlation of motion vectors. This method performs motion prediction of current block using the motion vectors of three spatially adjacent blocks and inter-frame prediction. Therefore the proposed method can predict motion precisely much more. Especially when the motion of objects is rapid, this method can estimate motion more precisely without reducing block size or increasing search area. Computer simulation results show that motion estimation of proposed method is more precise than that of conventional method.
분류 벡터 양자화 및 DPCM을 이용한 블럭 웨이브릿 변환 영상 부호화
김영춘,정원식,이건일,김경규,이승진 경북대학교 전자기술연구소 1997 電子技術硏究誌 Vol.18 No.1
In this paper, we propose a block wavelet transform coding of images using classified vector quantization to exploit the inter-band correlation and to eliminate the blurring effect at a smoothly changed edge. In this scheme, DWT coefficients are rearranged to compose the small blocks, which are made up of the corresponding coefficients from all the subbands. These blocks are classified into four classes relying on the directional energies of the second level subband along each direction to eliminate the blurring effect. These are further divided adaptively into subvectors depending on the DWT coefficient statistics as this allows efficient distribution of bits. The subvectors are then vector quantized and DPCM is applied to DC coefficients. Simulation results show that this coding scheme performs a good result compared to other approaches.
주변 블럭을 이용한 움직임 벡터에 대한 부가 정보 감소
김영춘,정원식,이건일 경북대학교 전자기술연구소 1995 電子技術硏究誌 Vol.16 No.2
In general, the motion vector of neighboring blocks would be similar. By using motion vector of neighboring blocks, side information to represent the motion vector for current block can be reduced. In this paper, we propose the method which reduce side information to represent the motion vector for each block using motion vector of neighboring blocks. When the motion vector for current block is equal to that for neighboring block, the position information of the same neighboring block instead of motion vector information is sent to the decoder. Therefore side information to represent the motion vector is reduced. By computer simulation, bit rates of proposed method is reduced to maximally 60.64% in comparison to that of conventional method.
평균 자승 오차와 발생 비트율을 이용한 적응 필드/프레임 선택 기법
김영춘,이건일,김중곤 경북대학교 전자기술연구소 1994 電子技術硏究誌 Vol.15 No.1
In this paper, a new adaptive field/frame mode selection method based on both the data rate and the minimum error is proposed. For each superblock, we decide the mode by the MSE and the data rate. If the modes decided by the MSE and the data rate criterion are equal, the mode information is sent to the decoder. On the other case, the final decision is made by the variation of bpp(bits per pixel) and variation of PSNR. The criterion has less MSE and bit rate, result in increase the compression efficiency.
김영춘,이건일,김중곤 경북대학교 전자기술연구소 1994 電子技術硏究誌 Vol.15 No.2
We show that the noise and impulsive components which deteriorate the spatial correlation are reduced by employing loop filter for motion compensated image(MCI). To reduce the bit rate of the interframe transform coder, a selective loop filtering method is proposed. The proposed method utilize the motion vector, variance, and generated bit rate to select a loop filter adequate to local image characteristics. When the motion vector is zero, no filtering method or fixed filtering method is selected based or the generated bit rate. And when the motion vector is non-zero, horizontally or vertically weighted loop filtering method is selected based on generated bit rate. An additional side information for the selection is transmitted to the receiver side.
김기완,오상광,이건일,마대영 경북대학교 전자기술연구소 1987 電子技術硏究誌 Vol.8 No.1
Photoconductive target for an image pick up tube was fabricated by thermal evaporation. For the good operation of the target it is necessary that the target have high photosensitivity, low lags and high resolution. In the vidicon mode analysis for the target, photosensitivity under 10 lux illumination was about 0.8 except the region of high intensity and the S/N ratio (dynamic range) was about 30㏈. The image lag was 20% at the 3rd field after shutting off light illumination. And the resolution measured with RETMA chart was about 300 TV lines. Considering all characteristics, the target was superior to those of Toshiba 20PE20.