Abstarct: The aim of geometrical explanation of Rock Mass Discontinuities is to careate a 3D model enabling us to define defining the basis for discontinuities. Often there is no access to underground level, in this situation the outcrops are used to gain information of engineering and structural characteristics of rock mass. In this research, image processing which is a safe, quick, low-cost and automated method is used to obtain geometrical parameters of Mean spacing, Density, Mean Trace Length (MTL) and fracture intensity. In this research the two algorithms of Hough Transform and Determining neighboring degree are developed to recognize the trace of discontinuities in quarry mines. In Hough Transform algorithm, first through applying median filter to the image, noises and microfracture are decreased and then using Canny Edge Detector, edges are revealed and short edges are removed by labeling connected components. Then, the lines indicating the lines of discontinuities are recognized using Hough Transform and the results from Hough transform are improved and Set Joints are recognized using Fuzzy-C Mean (FCM) and Subtractive Clustering. Ultimately, the results from image processing are compared with field results from Scanline Sampling, including scanline and circular scanline (Mauldon et al. Method) and window sampling, including window sampling, Square window sampling (Mauldon Method) and circular window sampling (Mauldon Method & Zhang & Einstein Method). In neighboring degree algorithm such as Hough Transform Algorithm, existing Noises and microfracture are decreased by applying filter median to image and then using Canny edge detector, existing edges are revealed and short edges are removed by labeling connected components. Thinning Operation was performed on the targeted picture and then the neighboring degree of lineament pixels was specified, the short lines are removed by labeling connected components afterwards, and discontinuity lines are defined according to neighboring degree. Finally, the separated lines are connected together.By comparing the results from different defined images with images with little lines and soft texture, both algorithms were able to determine traces well, however, neighboring degree algorithm presents better results as an outcome in complex images with higher number of traces.