Measuring Surfaces in Orthophotos Based in Color Segmentation Using K-means

Abstract:

This research presented a method to perform the measurement of orthophotography areas. An unsupervised segmentation was developed using K-means for each pixel to reconstruct an image with fewer colors to perform an area measurement through each given color tone. The Calinski-Harabasz index (CHI) was also studied; which allows a better choice of the color space in which the algorithm will be executed. The results of the execution times for each color space are presented, obtaining an error of approximately 0.08% in the measurement of the area.

Año de publicación:

2020

Keywords:

• K-Means
• Color spaces
• áreas
• Clustering
• Orthophotography

Fuente:

scopus