Nowadays, male infertility has increased worldwide. Therefore, a rigorous analysis of sperm cells is required to diagnose this problem. Currently, this analysis is performed based on the expert opinion. In order to support the experts in fertility diagnosis, several image processing techniques have been proposed. In this paper, we present an approach that combines the Lambertian model based on surface reflectance with mathematical morphology for sperm cells segmentation in micrographic images. We have applied our approach to a set of 73 images. The results of our approach have been evaluated based on ground truth segmentations and similarity indices, finding a high correlation between our results and manual segmentation.