Methods: The image analysis consists of two stages. These stages are measurement of parameters of desired comet images with automatic cell recognition and comet image classification. In MATLAB program, image preprocessing, segmentation and feature classification are applied. Otsu thresholding method is applied after the visualization of images containing comets with median and Gaussian filters. After these process, images in red channel of obtained binary images are obtained. Center of head are found by calculating the brightest 5% of comets on images. Density profile graph are obtained for each comet.

Results: There are 60 healthy and 25 damaged comets on five different RGB images. By the assistance of developed program, head percentage, tail percentage, length of head, length of tail, head area, tail area and whether they are healthy or damaged are defined on the images of comet. The success of the program in finding damaged comet was obtained as %85, and its success (sensitivity, specificity, accuracy) was found as %95.