由于成像质量不高，光照强度不均匀，皮下脂肪较厚等因素，近红外手肘静脉图像对比度较低，不易提取到清晰的静脉结构。针对该问题，本文提出了一种基于 Hessian算子的多尺度自适应 静脉滤波提取方法。该方法通过改进的多尺度自适应滤波器从对比度限制自适应直方图均衡化（Contrast-Limited Adaptive Histogram Equalization，CLAHE）增强后的图像中提取静脉。新的滤波器结构能够根据输入图像自适应地确定滤波器参数，在提取静脉的同时抑制噪声。实验结果 表明该方法可以有效地获得清晰完整的静脉结构，具有更强的去噪和增强效果以及更高的准确率。

A near-infrared elbow vein image has low contrast because of low image quality, uneven illumination intensity, and thicker subcutaneous fat. Therefore, it is difficult to extract a clear vein structure. To address this problem, a multiscale adaptive vein filtering enhancement method based on the Hessian operator is proposed. The method extracts veins from an image enhanced by contrast-limited adaptive histogram equalization with the use of an improved multiscale adaptive filter. The new filter structure can adaptively determine filter parameters based on the input image and suppress noise while extracting veins. The experimental results show that the method can effectively obtain a clear and complete vein structure, and it has stronger denoising, better enhancement effects, and higher accuracy.