Dermoscopic images allow to visualize with emphasis sections of the skin in which lesions are presumed. In these images it is common to find sections of body hair that represent noise in the analysis of the section for the treatment and diagnosis of skin cancer. In this work we present a proposal for the enhancement of the skin lesion of interest in dermoscopic images with abundant body hair over the lesion. The proposal is based on close-up visualization of the lesion of interest, by processing the image using a differential operator for edge detection, accompanied by a maximum variance threshold between classes for skin lesion detection. Tests were performed on images from the ISIC Challenge Dataset, and an average skin lesion size preservation rate of 93.852% and total body hair removal was determined for more than 80% of the images tested. The proposed method showed robustness to images of different sizes and pigmentation and can be considered as a support tool in computer-aided diagnosis processes for the treatment of skin lesions.