Abstract: The present disclosure is related in general to field of machine learning and image processing and a method and system for handling occluded regions in an image frame to generate surround view. A surround view generating device detects presence or absence of occluded blocks in each image frame received from image capturing devices associated with the vehicle, based on speed and direction of the vehicle. Further, occluded blocks are predicted by mapping the image frames and sensor data of the vehicle with pre-stored image data using a machine learning model. Finally, corrected image frames are generated by stitching the predicted occluded blocks to the image frames, and then a surround view of the vehicle is generated using corrected image frames and the plurality of image frames. The present disclosure enables prediction of occlusions caused due to dust or water depositions on the image capturing devices of non-overlapping field of view.

Abstract: The present disclosure is related in general to field of machine learning and image processing and a method and system for handling occluded regions in an image frame to generate surround view. A surround view generating device detects presence or absence of occluded blocks in each image frame received from image capturing devices associated with the vehicle, based on speed and direction of the vehicle. Further, occluded blocks are predicted by mapping the image frames and sensor data of the vehicle with pre-stored image data using a machine learning model. Finally, corrected image frames are generated by stitching the predicted occluded blocks to the image frames, and then a surround view of the vehicle is generated using corrected image frames and the plurality of image frames. The present disclosure enables prediction of occlusions caused due to dust or water depositions on the image capturing devices of non-overlapping field of view.