Change detection from Super-Resolution Imagery (SRI) plays a vital role in various fields, including environmental monitoring, agriculture, and urban planning. However, monitoring small-scale and complex land-use and land-cover changes remains a significant topic in the field of remote sensing. It is a considerable challenge due to limitations of the approach in obtaining both spatial and temporal changes effectively. To address this challenge, a deep learning-assisted framework, the Vision Transformer-based Dual Metric Deep Learning Technique (ViT-DMDLT), is proposed for detecting small changes in super-resolution satellite images. The proposed framework leverages the strength of Vision Transformers (ViTs), Convolutional Neural Networks (CNNs), and the Dual Metric Network (DMN). Moreover, the integration of low-resolution and high-resolution satellite images is carried out to obtain the super-resolution satellite images. Furthermore, the ViT encoders are utilized to capture the overall spatial dependencies, and a dual metric ensures robust features by reducing the intra-class changes and enhancing the inter-class separability among the temporal images. The effectiveness of the proposed framework, ViT-DMDLT, is validated on three different publicly available datasets named SYSU, Cropland, and LEVIR-CD. The outcome of the framework has achieved superior performance with an overall accuracy of 96.78% on SYSU, 90.77% on Cropland, and 98.12% on LEVIR-CD. The results of the framework have shown robustness compared to the other state-of-the-art models. These outcomes highlight the effectiveness of the ViT-DMDLT framework by detecting the land cover changes with high accuracy, even in complex and small variations from super-resolution images.