Today by the increasingly growth of internet users, many internet services such as electronic mails, search engines and social networks are presented. Due to developments in internet services, disabling the services through automatic programs such as internet robots and hackers has become a main problem. CAPTCHA is a mechanism that confirms whether the user is a human or a computer program. This task is performed through some tests. Performing these tests is easy for human but it is difficult for robots. CAPTCHA techniques are trying to increase the mentioned difficulties in automatic programs and it is performed through increasing the deviations and parasite. Then solving it will become difficult for human. In this research an image CHAPTCHA is presented and it is done through mathematical morphology the purpose of this solution to increase security and to defend against malicious attacks and hackers. In this method by the input image, a binary model is developed and worn. Then by making some changes, the image is presented to the user and he should recognize the early image. According to these estimations, 96% human have passed the face CAPTCHA and 90.5% human have passed the non-face CAPTCHA, and the average of required time to answer is 16.8 for non-face CAPTCHA and 7.6 second for face CAPTCHA.