Principles of Image Pyramid

An image pyramid is a collection of images generated according to certain rules, with resolutions ranging from fine to coarse. The image pyramid technology uses image resampling methods to create a series of image layers with different resolutions. Each layer is segmented and stored, and a corresponding spatial index mechanism is established, thereby improving the display speed when zooming and browsing images. As shown in the image pyramid below, the bottom represents the original highest resolution of the image, with an image resolution of 512×512. The higher up the image, the smaller the resolution, which are 256×256 and 128×128 respectively. The top is the lowest resolution image of the image pyramid, 64×64. Therefore, this image pyramid has four layers, i.e., four levels of resolution. Obviously, the higher the image resolution, the more levels the image pyramid has. For an image with a resolution of 2^a×2^b (a>b), SuperMap will create an image pyramid with (b-6)+1 layers.

Figure: Image Pyramid

After building an image pyramid for the image, each time you browse the image, the system will retrieve its image pyramid to display the data. When you zoom in or out the image, the system automatically selects the most appropriate pyramid level based on the user's display scale to display the image.

Pyramids can only be created for original data. You can only build pyramids for one dataset at a time. If you want to create it again, you need to delete the already created pyramid. When browsing a raster dataset after creating an image pyramid, you are actually accessing the created pyramid. The process of building pyramids at different scales is shown in the figure below.

Creating an image pyramid can significantly improve the speed and performance of image zoom display rendering. However, at the same time, building an image pyramid will increase the storage space of the dataset, i.e., increase the size of the datasource file where the image dataset is located. This is because the built image pyramid is essentially a collection of images at different resolutions, and these images at different resolutions are stored together with the data in the datasource file, thereby increasing the size of the datasource file. Moreover, the larger the amount of raster dataset data, the longer the time required to build the pyramid, and the larger the storage space for the image pyramid. However, it will save more time for future image browsing. Therefore, for massive image data, building pyramids is a good choice for optimizing efficiency.

The SuperMap SIT image file (obsolete) storage format is a data format that integrates image compression and efficient image pyramid technology, thus enabling ultra-fast display of image data, largely independent of the size of the image data. Even on machines with very low configurations, it can smoothly display massive image data.

Related Topics