Feature point detection method and device and computer readable storage medium

Abstract

The invention discloses a feature point detection method, a feature point detection device and a computer readable storage medium. The detection method comprises the following steps: dividing a to-be-processed image into a plurality of image blocks; on the to-be-processed image, anchor frames are generated with the preset points in the image blocks as centers; respectively predicting the offset of each anchor frame relative to the target feature point and the confidence of each anchor frame relative to the target feature point; and determining the target feature point according to the position of each anchor frame, the offset of each anchor frame relative to the target feature point and the confidence of each anchor frame relative to the target feature point. The detection method provided by the invention can improve the accuracy of feature point detection.

Description

technical field [0001] The present application relates to the technical field of feature point detection, and in particular to a feature point detection method, detection device, and computer-readable storage medium. Background technique [0002] When the traditional algorithm detects the feature points, it adopts the method of full-image regression. The specific method is: first calculate the proportion of the feature points on the whole image, and then multiply the width and height of the image to obtain the position of the feature points. The method makes the error of the feature point position related to the size of the image. The larger the image, the greater the error. Especially in a large-scale image, the error of the feature point may have 10 to 20 pixels, which cannot meet the requirements of subsequent industrial positioning. Contents of the invention [0003] In view of this, the present application provides a feature point detection method, a detection device,...

AI Technical Summary

Problems solved by technology

[0002] When the traditional algorithm detects the feature points, it adopts the method of full-image regression. The specific method is: first calculate the proportion of the feature points on the whole image, and then multiply the width and height of the image to obtain the position of the feature points. The method makes the error of the feature point position related to the size of the image. The larger the image, the greater the error. Especially in large-size images, the error of the feature point may have 10 to 20 pixels, which cannot meet the requirements of subsequent industrial positioning.

Images