Collecting updated and accurate land cover changes in urban areas have a significant impact on urban planning and management. In recent decades, remote sensing data have been used as valuable sources for detecting land cover changes. Given the different information that optical and radar sensors have received from any phenomenon on earth's surface, remote sensing data are assumed as complementary tools, and the integration of these two kinds of data will improve the results in detecting changes, especially in urban areas. In this research, a non-supervised and automatic method was developed to improve the detection of land cover changes in urban areas by integrating radar and optics data. Different spectral indices and radar polarizations were used to develop the CVA technique, known as an efficient non-supervised method for detecting the variations. In the implementation section, Sentinel 1 and 2 satellite data were used for the period of 2106 to 2018, captured from the northwest of Mashhad city, Iran. The developed technique was compared with other change detection methods. The findings of this study indicated the effectiveness and accuracy of the developed technique for detecting the changes. The estimated ratio of detected pixels to total pixels was 82%, which was promising. The overall classification accuracy and the kappa coefficient with values of 90.17 and 0.8016 were highest among the other methods used in the present study. The non-supervised approach and the verification results of the proposed method revealed its usefulness in detecting the changes, especially in urban areas.