Method for judging small-scale motion direction of image

The invention discloses a method for judging the small-scale motion direction of an image, and the method comprises the steps: firstly carrying out the partitioning of a blurred image with a smaller motion scale; secondly, performing super-resolution reconstruction on the selected image blocks by using an SRCNN network, and amplifying a linear smear effect caused by motion blur to the image; and finally, using the constructed angle differential filter to calculate the gradient of the block in each direction for the super-resolution reconstructed block. In order to solve the problem that the blurring direction is difficult to identify due to the fact that the motion scale of an image is too small, such as 3-10 pixels, a partitioning method, a block selection method, a block super-resolution reconstruction method and a weighted average angle differential filtering method are sequentially adopted. Aiming at a small-scale motion blurred image, the method is relatively high in angle identification precision, relat