Abstract: Image sharpening is a method which makes the digital image much more clearly by compensating outlines and sharpening edges of image.When normal sharpening algorithm performs high frequency enhancement，the obvious noises are also enhanced.Aiming at the detects，a non-linear image sharpening approach is proposed based on noise estimation which can be controlled.Because the details and noises of the image are difficult to distinguish，the algorithm can control the parameter to divide the noise from the edge.Experiment results demonstrate that this algorithm can not only sharpen the edge，but also suppress the noise，this algorithm enhances the image detail，the values of PSNR and DV/BV are higher，moreover the image enhanced has good visual effect.

Key words: non-linear sharpening algorithm, detail enhancement, noise estimation

摘要： 图像锐化是一种补偿轮廓，突出边缘信息以使图像更为清晰的处理方法。常规的锐化算法对图像进行高频增强时，结果会呈现明显噪声，因此提出了一种基于噪声控制的非线性锐化增强方法。针对图像边缘细节和噪声难以区别对待的缺点，该算法通过控制非线性函数中的参数来有效地将噪声和边缘分开处理。实验证明，该方法不仅能够锐化图像边缘，而且改善了传统锐化算法对图像噪声放大的缺点，经算法处理后的图像细节丰富，峰值信噪比和DV/BV值较高，具有良好的视觉效果，很好地改善了图像质量。

关键词: 非线性锐化算法, 细节增强, 噪声限制