Abstarct: Our surrounding environment is a 3 dimensional space and all the existing objects in this environment have length, width and height (depth). When we are looking at this environment or taking photos with typical cameras, actually we are mapping a 3 dimensional space into a 2 dimensional one on the paper. This process can be ended up to losing the third dimension which is depth. Depth has some critical information in order to analyze position and features of the objects in the picture. Human brain can understand the environment objects in a 3 dimensional way using brain 3 dimensional reconstruction techniques and binocular vision ability while we should reconstruct the third dimension in photos taken from objects using available computational techniques. Image depth estimation is an important issue in machine vision and artificial intelligence. However researchers have not achieved certain solutions for this problem. The main purposes of this thesis are to calculate the depth of a certain scene using its 2 dimensional monocular image and depth wise color segmentation of the image. To reach this goal in first step, some specific features have been extracted from the image and formed the feature vector. As the second step, this vector would be fed to a neural network. The output of the network, image can be divided into distinct classes according to their depth. In the third step, each class will be colored baxsed on its depth. The proposed method in this thesis has some advantages in comparison with the prior methods including, Independence of camera and scene specifications, high precision and speed, no need to image classification and getting typical and simple techniques of image processing into work.