Abstract: In view of the question of image quality assessment method with no reference at the present, such as lower quantitative accuracy and affected by the image contrast, texture and structure, it is hard to used in the detection of high-throughput sequencing image quality. To analyze the white on black background image superposition rule shapes the homomorphism of Fourier transform, a model of image quality assessment on frequency domain is presented, the model bases on frequency domain of image texture structure pretreatment, and finally realizes the high throughput sequencing image quality evaluation index. Practice shows that the method has a better quantitative precision and reliability when it is used to evaluate the high-throughput sequencing image quality.

Key words: high-throughput genetic sequencing image, image quality assessment, frequency domain, texture structure, homomorphic Fourier

摘要： 目前无参考图像质量评价方法中存在的量化精度不高和受图像对比度、纹理结构影响大等不足，难以用于高通量基因测序图像质量的检测。为此，在分析了黑色背景图像上叠加规则白色形状的同态傅里叶变换的基础上，提出了一种基于频率域的图像质量评价模型。该模型通过对频率域所表现出的图像纹理结构进行预处理，最终建模实现了高通量基因测序图像质量的评价指标。实践表明，该方法在评价高通量基因测序图像质量中具有较好的量化精度和可靠性。

关键词: 高通量基因测序图像, 图像质量, 频率域, 纹理结构, 同态傅里叶