Synthetic Aperture Radar (SAR) remote sensing plays an important role in research areas such as resource management, forest conservation, crop monitoring, land hazards monitoring, elevation product generation, and military applications. SAR has active imaging capability with an ability to discriminate terrain features, along with recognition of selected natural as well as man-made targets. However, special abilities of SAR become ineffective in specific cases due to interference of SAR frequency bands with the same magnitude range of radio frequencies originating from other types of electronic equipment. This equipment may include air-traffic surveillance radars, meteorological radars, communication systems, Radio Local Area Network (RLAN), and other electromagnetic (EM) radiation sources. This process of SAR frequency band contamination is called Radio Frequency Interference (RFI). Due to the increasing communication applications based on EM radiation, a wide range of EM spectrum is being used for this purpose. SAR frequency bands are very closely packed and even overlapping with other operating frequency bands allotted for other applications. Due to gaps in the unified international planning for EM spectrum band allocation for different applications, the problem of RFI in every communication application is rising rapidly. The satellites of the Sentinel-1 constellation use a radar, which operates in the IEEE (Institute of Electrical and Electronics Engineers) standard defined C band (central frequency 5.405 GHz) which covers most civilian and defense use frequencies. The RFIs discussed in the study manifest themselves on Sentinel-1 data in the form of lines having bright signatures, which are always perpendicular to the satellite orbit trajectory. These patterns may be hundreds of kilometers long and signify that a powerful radio source close to 5.405 GHz (such as some radars) is active and emitting somewhere along those lines. These interference patterns rigorously reduce the SAR image quality, which results in reducing the usefulness of SAR images, especially for high-resolution data-based applications. Therefore, an effective RFI pattern detection method is necessary for prior identification of RFI contaminated SAR images. In this study, openly accessible Sentinel-1 dual polarimetric (GRD) SAR images taken over different busy maritime shipping ports having international trade such as in Dubai and Germany have been used for the semantic segmentation of RFI patterns. The RGB composite images of different experimental sites were used to test and train the U-Net architecture of Convolutional Neural Network (CNN) for RFI pattern recognition.