Diabetic Retinopathy treatment or detection requires the extraction of blood vessels from the Fundus image, which can be further used in the feature extraction process, and by using these, an automated system can be developed for early diagnosis of the disease. This paper has tried to implement an algorithm that can be used to detect the blood vessels of the retina from Fundus images. The algorithm has shown some promising results on the dataset taken from IDRID. The database has more than 750 Fundus images categorized into seven stages of the disease, taken under different lighting conditions.