Retinal image blood vessel segmentation system based on multiple detrending analysis

The invention discloses a retinal image blood vessel segmentation system based on multiple detrending analysis. The system includes an image preprocessing unit for performing green channel extractionon the original image and acquiring the target region and performing enhancement processing on the target region image, and adopting multi-fractal detrending analysis. Methods A blood vessel segmentation unit for extracting retinal blood vessel images and a retinal image post-processing unit for linking the extracted retinal blood vessel images with broken segmentation and removing isolated pixelswere performed. The invention extracts blood vessels from the retinal images of lesions, adopts multiple detrending analysis algorithms that are good at describing complex situations, realizes accurate extraction of lesion image blood vessels, and improves the clinical practicality of the system and accuracy of blood vessel segmentation for special conditions. 本发明公开了种基于多重去趋势分析的视网膜图像血管分割系统，系统包括对原始图像进行绿色通道提取获取目标区域并对目标区域图像进行