Vein Extraction Method Based on Improved Canny Operator Based on Window Dynamic Threshold

Abstract

The invention discloses a leaf vein extraction method based on a window dynamic threshold improved Canny operator, which includes correcting the leaf angle, using a bilateral filter to smooth and correct the leaf image, extracting the main leaf vein, using a bidirectional region growing method to remove noise and extracting the leaf vein, etc. The present invention uses bilateral filtering to replace the Gaussian filtering used in the traditional Canny operator, so that more edges are retained in the result of subsequent edge extraction, and the connectivity of long edges is improved; the Canny operator based on window dynamic threshold improvement is used Instead of the traditional Canny operator with a global fixed Canny threshold, it effectively complements the missing veins; using the bidirectional region growing method, it extracts and separates the primary and secondary veins with relatively complete details while removing noise.

Description

technical field [0001] The invention relates to a leaf vein extraction method based on a Canny operator, in particular to a leaf vein extraction method based on an improved Canny operator based on a window dynamic threshold, and belongs to the technical field of computer vision. Background technique [0002] Vein extraction refers to extracting the pixels that make up the primary and secondary veins of the leaf from the vein images of different shapes and placements by means of computer vision. Existing methods mainly include cluster classification, edge detection, neural network extraction and so on. It is of great significance for the identification of leaf species and the construction of leaf vein network system. [0003] The Canny operator is an edge extraction operator. For a grayscale image, by setting a double threshold and using the first-order difference seed point to expand, the edge of the image is finally obtained. The specific steps include Gaussian filtering,...

AI Technical Summary

Problems solved by technology

[0004] For the Canny operator, if the appropriate threshold can be selected for different scenes, the edge with a good signal-to-noise ratio can be extracted. However, the setting of the double threshold often needs to be manually adjusted through experience; at the same time, the There are sometimes large differences in the edge strength of each local area. In the traditional Canny operator, the double threshold is shared globally, so it is often difficult to obtain a threshold suitable for all areas.

As a result, when it is used for vein extraction, a large number of vein fracture areas and non-vein edges are generated, and different thresholds need to be set for different leaves. Therefore, an effective method of giving different edge strengths different double thresholds is needed.

Images