Improved kernel density peak clustering method for plant image segmentation

Abstract

The invention provides an improved kernel density peak value clustering method for plant image segmentation, and the method comprises the steps: selecting a central point through employing a decision diagram, mapping data to a high-dimensional space through employing an RBF, and carrying out the clustering of sample points in the high-dimensional space through employing two distribution strategies, therefore, the defects of the K-Means clustering algorithm and the DPC clustering algorithm in agricultural image segmentation are better overcome. Experiments are carried out on a plurality of artificial data sets and UCI data sets, comparison with other clustering algorithms is carried out, finally the clustering algorithms are applied to plant image segmentation, and experimental results show that the algorithm of the invention has a good clustering effect.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an improved kernel density peak clustering method for plant image segmentation. Background technique [0002] When studying crops, plants require additional water and fertilizer under different conditions, so it is necessary to know exactly how the current crop is growing. Image segmentation is the basic technology of image processing, and the application demand of this technology in agriculture is getting stronger and stronger. With the deepening of the research on image segmentation technology, many machine learning methods have also been applied in this field. The purpose of clustering is to divide similar points into similar groups, and it has been widely used in many fields. The K-means method is a distance-based clustering method with fast convergence speed and robust generality. It is one of the more effective methods among many clustering methods. In t...

AI Technical Summary

Problems solved by technology

However, as the assignment proceeds, if a point is mistakenly assigned to an irrelevant, incorrect cluster, a chain reaction will result in a series of points being incorrectly assigned to irrelevant clusters

(2) For some nonlinear data sets, especially in data sets with higher dimensions or data sets with overlapping point distributions, the DPC method cannot achieve satisfactory clustering results

Images