Remote sensing change detection is a rapidly emerging field with applications in many different domains, including environmental monitoring, disaster monitoring, and the identification of alterations in land cover and land use. This study deals with a bi-temporal image transformer-based model for change detection in remote sensing. The method uses a pair of images of a place at two different time periods to forecast the change mask. These images will be first given to a ResNet CNN backbone network for feature extraction. These extracted features will be propagated to a tokenizer network to extract context-rich tokens from the extracted feature maps. Later these tokens will be concatenated and given to the transformer encoder network. In this study, the transformer employed is a stack of encoders and a stack of decoders network with a ResNet backbone. The results from the encoder network are propagated to the decoder network. From the experimental setup, it was clear that the current model showed higher performance compared to the existing models FC-EF, FC-Siam-Conc, FC-Siam-Di, IFN, BiT, and HANet with a precision of 92.04, recall of 90.27, MIoU of 84.05, and an mF1 of 91.06.