作者: Hong Kwang Seok , Oh Byung Hun
DOI:
关键词:
摘要: A method (200) for detecting a color object in an image includes step (210) to receive by processing unit; (220) of performing RGB normalization with respect the image; (230) converting normalized into YCbCgCr model; and (240) extracting having value greater than or equal threshold chrominance channel each Cb, Cg Cr from model. Further, further (250) maximum specific information corrected Cr. Thereafter, may include (260) binarization model (270) remove noise binarized using at least one erosion operation dilation operation.