发明名称 Image processing method
摘要 This invention provides an image processing method which for noise reduction and sensitization for an ordinary video camera. The noise reduction processing locates a bright pixel and averages brightness by adding the pixel accumulated by ratio in accordance with the geometric series, and to a dark pixel, carrying out the processing of noise reduction which averages brightness by adding the pixel accumulated by ratio in accordance with the geometric series and the processing of sensitization in the condition that a magnification of intensification is greater than 1, determine that the pixel moves or not, if the pixel is moving, it is used to carry out processing of sensitization only, and if the pixel is still, choose it is used to carry out the processing of sensitization and noise reduction.
申请公布号 US8780235(B2) 申请公布日期 2014.07.15
申请号 US201213517569 申请日期 2012.06.13
申请人 Logic & Systems, Inc. 发明人 Kobayashi Masahiro
分类号 H04N5/217;H04N9/64;H04N5/76;G06K9/40 主分类号 H04N5/217
代理机构 代理人 Sharra Lee Fredric
主权项 1. An image processing method to process animation data photographed with video cameras, this image processing method can carry out processing of noise reduction/sensitization and processing of movement detection for every pixel in parallel; the processing of noise reduction/sensitization is comprised of two steps of a first and second processing, in the first processing, it does not carry out the processing of sensitization to a bright pixel, and carries out the processing of sensitization to a dark pixel in the condition that a magnification of intensification is greater than 1, in the second processing, it carries out the processing of noise reduction which averages brightness by adding the pixel accumulated by ratio in accordance with the geometric series to a bright pixel, and carries out the processing of noise reduction which averages brightness by adding the pixel accumulated by ratio in accordance with the geometric series and the processing of sensitization in the condition that a magnification of intensification is greater than 1 to a dark pixel, regarding to the processing of movement detection, store the brightness of the latest and a past plurality of frames on which the processing of noise reduction and the processing of sensitization are not carried out in the first ring buffer, store a plurality of the total value of the brightness of the latest and a past plurality of frames in the second ring buffer, calculates the absolute value of the difference between the total value of the brightness of the old frames and the total value of the brightness of the latest frames, those are memorized in the second ring buffer using the first calculating means, and determine that the pixel moves when the absolute value is larger than a constant value, regarding the pixel which is judged as moving by the processing of movement detection, choose the pixel value which was obtained by the first processing of the processing of noise reduction/sensitization, and for the pixel which is judged as not moving by the processing of movement detection, choose the pixel value which is obtained by the second processing of the processing of noise reduction/sensitization, and these chosen values output as a video image.
地址 Kobe JP