Agricultural Super Green Image Segmentation Method Based on PSO and K-means Algorithm
The main purpose of this paper is to study the method of agricultural super green image segmentation based on
PSO and K-means algorithm. This paper mainly introduces the research background, purpose and significance
of this topic, as well as the research status at home and abroad, and expounds the theoretical basis of image
acquisition and image segmentation. After the introduction of clustering analysis and particle swarm
optimization algorithm, an improved image segmentation method based on PSO-K algorithm is proposed. Corn,
weed and apple were selected for the experiment of agricultural super green image segmentation and the quality
of image segmentation was analyzed. The experimental results show that the NRECR index quality of maize
image using only k-means algorithm is 0.116, and that of maize image using only PSO algorithm is 0.935.
However, the NRECR index quality of maize image using PSO-K improved algorithm proposed in this paper is
as high as 0.942. The NRECR index quality of weed image only using k-means algorithm is 0.432, which is 0.5
lower than that of weed image only using PSO algorithm or only using PSO-K improved algorithm. PSO-K
improved algorithm solves the problems of slow convergence speed of particle swarm optimization algorithm
and the correlation between k-means algorithm and initial clustering center.