“In response to the high demand for hardware resources in the field of remote sensing image super-resolution reconstruction, the research team has proposed an innovative lightweight network architecture. This architecture utilizes reparameterization techniques to design a residual local feature module, effectively enhancing the ability to extract local features from images. At the same time, a lightweight global context module has been introduced, which can associate similar features within the image, thereby enhancing the network's feature expression ability. In addition, by adjusting the channel compression factor of the module, the research team successfully reduced the number of model parameters and improved model performance. The test results on the UC Merced remote sensing dataset show that the parameter count of this algorithm is only 539K at 3x super-resolution of remote sensing images, far lower than the 5470K of the HSENet algorithm. At the same time, the algorithm also shows advantages in peak signal-to-noise ratio and structural similarity, reaching 30.01 dB and 0.8449, respectively, surpassing the 30.00 dB and 0.8420 of the HSENet algorithm. In terms of inference speed, this algorithm also shows significant advantages, requiring only 0.010 seconds, while the HSENet algorithm requires 0.059 seconds. In addition, testing on the DIV2K natural image dataset further validates the algorithm's generalization ability, and its peak signal-to-noise ratio and structural similarity also demonstrate certain advantages compared to other algorithms. This research achievement provides a new solution for the field of remote sensing image super-resolution reconstruction, which is expected to promote technological progress in this field.”