Abstarct: Reservoirs in the middle east are among the most important fractured reservoirs in the world such as nature and degree of folding, fault - corrosion, in - situ stresses and changing characteristics of rock such as porosity, laxyering, lithology and shale percent as well as the main role in permeability estimation and consequently their productivity. Image log tools are one of the first sources of information for the study of rift reservoirs, enabling geologists to identify significant details on the lithological status and correlation between stratigraphic structure, groundwater flow status, fracture identification, faults, laxyers and veins within the rock laxyers, as well as facilitating the identification of lava flows for mining and geotechnical assessments, require the use of multi-sensor sampling tools to obtain such information. One of the most powerful tools for studying and identifying fractures around the wells is the FMI image logs. These tools produce an indirect, high-resolution image of the well. The purpose of this study was to use Canny edge detection algorithm to identify sinusoidal edges of fractures from FMI log data in two wells A and B of one of the southwestern hydrocarbon fields of Iran. Fractures in these two wells were studied in two ways. In the first method, by integrating the results of the Geolog interpretation software manually, open and closed fractures were identified and separated with high precision. Finally, fractures were identified in well A with a total of 264 open and closed fractures with a general trend of north-east-southwest and in well B with a total of 60 open and closed fractures with a general trend of east-west and were determined by comparing fracture density and fracture openings in each. Two wells with the highest fracture and opening density were also identified. In the second method, considering the fractures in the FMI images as an edge, the Canny filter, one of the most powerful edge recognition filters in image processing, was used to evaluate the performance of this filter on FMI images. In order to perform the Canny algorithm for fracture detection, the Sobel algorithm, another edge-baxsed algorithm, was used to compare these two methods. Comparison of the results of this method with the results of the fracture detection using Geolog software shows that the use of Canny edge detection algorithm is very good in determining fractures and can help the interpreter to detect fractures on the FMI image log.