An image enlarging method, including the following steps: first, divides an image into several sampling sub-area; next, decides the reference value of each sampling sub-area; then, compares the reference value and a critical value to get a comparing result; final, computes the pixel value that needs to fill. The method of the invention includes: divides an image into several sampling sub-area; decides the high frequency heft of each sampling sub-area; perform a first arithmetic according to the high frequency heft; perform a second arithmetic according to the high frequency heft. The invention can use the sampling sub-area appropriately to choose the fit arithmetic, and switch different arithmetic according to the different image characteristic, set the critical value by itself according to the actual need of the deviser and the user. |