提出了一种基于Beamlet变换的算法,克服了图像中的各种干扰及噪声,试验结果表明其具有较好的实用价值。
An algorithm based on Beamlet Transform is proposed. Having dyadic and recursire subdivision on image, intgrating on Beamlets in each subsquare, the straight lines are approximated under GLRT(Generalized Likelihood Ratio Testing). This method can overcome all kinds of noise in images and can extract exactly straight lines even in low SNR images.