적외선 영상의 동적 범위 압축을 위한 적응적 지역 히스토그램 변형 기법 (Adaptive local histogram modification method for dynamic range compression of infrared images)

서지정보

· 발행기관 : 한국컴퓨터정보학회
· 수록지 정보 : 한국컴퓨터정보학회논문지 / 24권 / 6호 / 73 ~ 80페이지
· 저자명 : 정지혜

초록

In this paper, we propose an effective dynamic range compression (DRC) method of infrared images. A histogram of infrared images has narrow dynamic range compared to visible images.
Hence, it is important to apply the effective DRC algorithm for high performance of an infrared image analysis. The proposed algorithm for high dynamic range divides an infrared image into the overlapped blocks and calculates Shannon’s entropy of overlapped blocks. After that, we classify each block according to the value of entropy and apply adaptive histogram modification method each overlapped block. We make an intensity mapping function through result of the adaptive histogram modification method which is using standard-deviation and maximum value of histogram of classified blocks. Lastly, in order to reduce block artifact, we apply hanning window to the overlapped blocks.
In experimental result, the proposed method showed better performance of dynamic range compression compared to previous algorithms.

영어초록

In this paper, we propose an effective dynamic range compression (DRC) method of infrared images. A histogram of infrared images has narrow dynamic range compared to visible images.
Hence, it is important to apply the effective DRC algorithm for high performance of an infrared image analysis. The proposed algorithm for high dynamic range divides an infrared image into the overlapped blocks and calculates Shannon’s entropy of overlapped blocks. After that, we classify each block according to the value of entropy and apply adaptive histogram modification method each overlapped block. We make an intensity mapping function through result of the adaptive histogram modification method which is using standard-deviation and maximum value of histogram of classified blocks. Lastly, in order to reduce block artifact, we apply hanning window to the overlapped blocks.
In experimental result, the proposed method showed better performance of dynamic range compression compared to previous algorithms.

참고자료

· 없음