A image defect detecting vision system for the automatic optical inspection of wafer has been developed. For the successful detection of various kinds of defects, the performance of two threshold selection methods are compared and the improved Otsu method is adopted so that it can handle both unimodal and bimodal distributions of the histogram equally well. An automatic defect detection software for practical use was developed with the function of detection of ROI, fast thresholding and area segmentation. Finally each defect pattern in the wafer is classified and grouped into one of user-defined defect categories and more than 14 test wafer samples are tested for the evaluation of detection and classification accuracy in the inspection system.

The purpose of the study was to evaluation of the radiation dose reduction using various automatic exposure control (AEC) systems in different manufactures multi-detector computed tomography (MDCT). We used three different manufacturers for the study: General Electric Healthcare, Philips Medical systems and Siemens Medical Solutions. The general scanning protocol was created for the each examination with the same scanning parameters as many as possible. In the various AEC systems, the evaluation of reduced-dose was evaluated by comparing to fixed mAs with using body phantom. Finally, when we applied to AEC for three manufacturers, the radiation dose reduction decreased each 35.3% in the GE, 58.2% in the Philips, and 48.6% in the Siemens. This applies to variety of the AEC systems which will be very useful to reduce the dose and to maintain the high quality.

스마트 폰 사용자의 수가 점차적으로 증가함에 따라 모바일을 이용한 식물정보 관련 애플리케이션서비스가 필요 되어 지고 있다. 이에 본 연구는 식물정보 서비스 애플리케이션 개발에 필수 요소인 식물인식 및 분류 전산화 과정의 알고리즘을 제안하였다. 연구를 통해 독자적으로 개발한 Sweep(SP) 외곽선 추출 알고리즘을 이용 하여 외곽선을 검출하고 형태적인 특징요소를 정의하였으며 검출 된 외곽선을 이용하여 H/W ratio, Top tip ratio, Bottom tip ratio, 등분각 연장선과의 교차점 위차 정보, 등분각 연장선과의 교차점 거리 정보, 근접이차함수 비교 등 총 6가지 분류 기준을 전산화하였 다. 제안한 분류 기준의 유효성 검증을 위하여 총 32종의 식물을 재 료로 실험한 결과 H/W ratio과 Top tip ratio는 식물별 고유한 특성 을 기준으로 소수 그룹을 형성하고 구분할 수 있는 유효성을 가지 는 것으로 검증되었다. 등분각 연장선과의 교차점을 이용한 거리정 보는 식물 종류 간의 패턴 차이가 인정되었으며 이의 전산화를 위 하여 바타차야 비교연산법을 적용하여 상대비교 알고리즘을 완성 하였다. 또한 본 연구 과정에서 의도하지 않았던 엽저 및 엽선의 경 향 그리고 결각의 유무 검출 가능성을 확인하게 되어 추가적인 알 고리즘 개발이 가능할 것으로 여겨진다.

This study was conducted to investigate the reliability of automatic cracked and bloody egg detector according to the age of the hens and the level of the detector. The results of this study are expected to be helpful in the implementation of the Korean egg grading system, which is expected to improve egg quality for consumers. An official egg grader randomly selected 1,000 eggs for each experiment (total 36,000 eggs), ran them through the automatic detector, and conducted labor inspection using the eggs that were classified by the detector as cracked, bloody, and normal eggs. The results showed that more cracked eggs were laid by hens aged 40-60 weeks than by hens aged 30 weeks (p<0.05). Also, when the detector level increased from four to seven (i.e., when it became less sensitive), its cracked eggs detection rate dropped, and the total rate of cracked eggs was consistent after the labor inspection of the classified eggs. The automatic detector achieved over 97 percent accuracy. The bloody eggs constituted only 0.005 percent of all the samples, and all the detector-detected eggs were bloody eggs after the labor inspection of both the bloody and normal egg lines. Therefore, it can be concluded that the automatic cracked and bloody egg detector was reliable and can be used in the egg grading system. Considering that cracked eggs should be less than 9 percent of first-grade eggs in the present egg grading system, the use of an automatic crack detector may help provide better-quality eggs to consumers by producing less than 5.5 percent cracked eggs.