This paper introduces a multiresolution spline technique for combining images into a larger image mosaic. The method involves decomposing images into band-pass filtered components, assembling these components into band-pass mosaics, and then summing the mosaics to form the final image. The technique ensures that the spline matches the scale of features in the images, allowing for smooth transitions between images without blurring finer details near borders. The key idea is to use a pyramid algorithm to decompose images into multiple frequency bands. Each band is then processed with a spline that adjusts the transition between images to ensure smoothness. The multiresolution approach allows for handling images with varying spatial frequency content, ensuring that both large-scale and small-scale features are properly blended. The technique is applied to various image mosaics, including synthetic star images, Landsat images of San Francisco, and images of an apple and an orange. The results demonstrate that the multiresolution spline effectively eliminates visible seams between images, even when the images are very different or have irregular shapes. The pyramid structure is particularly well-suited for this task, as it allows for efficient filtering and splining operations. The algorithm uses a set of low-pass filtered images, which are then combined to form band-pass images. These band-pass images are then splined to create the final mosaic. The method is efficient, requiring only a small number of arithmetic operations per pixel. The weighting functions used in the splining process are implicitly defined by the pyramid computation, ensuring that they are matched to the frequency bands represented in the pyramid. This approach avoids artifacts such as blurred edges and double exposure effects that can occur with simpler weighted average techniques. Overall, the multiresolution spline provides a practical and general technique for forming image mosaics, combining the benefits of multiresolution analysis with efficient filtering and splining operations. The pyramid structure offers a unifying framework for both filtering and splining, making it a powerful tool for image processing applications.This paper introduces a multiresolution spline technique for combining images into a larger image mosaic. The method involves decomposing images into band-pass filtered components, assembling these components into band-pass mosaics, and then summing the mosaics to form the final image. The technique ensures that the spline matches the scale of features in the images, allowing for smooth transitions between images without blurring finer details near borders. The key idea is to use a pyramid algorithm to decompose images into multiple frequency bands. Each band is then processed with a spline that adjusts the transition between images to ensure smoothness. The multiresolution approach allows for handling images with varying spatial frequency content, ensuring that both large-scale and small-scale features are properly blended. The technique is applied to various image mosaics, including synthetic star images, Landsat images of San Francisco, and images of an apple and an orange. The results demonstrate that the multiresolution spline effectively eliminates visible seams between images, even when the images are very different or have irregular shapes. The pyramid structure is particularly well-suited for this task, as it allows for efficient filtering and splining operations. The algorithm uses a set of low-pass filtered images, which are then combined to form band-pass images. These band-pass images are then splined to create the final mosaic. The method is efficient, requiring only a small number of arithmetic operations per pixel. The weighting functions used in the splining process are implicitly defined by the pyramid computation, ensuring that they are matched to the frequency bands represented in the pyramid. This approach avoids artifacts such as blurred edges and double exposure effects that can occur with simpler weighted average techniques. Overall, the multiresolution spline provides a practical and general technique for forming image mosaics, combining the benefits of multiresolution analysis with efficient filtering and splining operations. The pyramid structure offers a unifying framework for both filtering and splining, making it a powerful tool for image processing applications.