Remote Sensing Technology and Application, Volume. 40, Issue 2, 461(2025)
Methods of Improving Land Cover Classification based on Terrain Factors and Time-series NDVI
Land cover types and their distribution are closely related to terrain factors and time-series NDVI. For research of land cover classification accuracy, terrain factors and time-series NDVI were used to improve the accuracy of land cover classification for the first time. In this study, the MODIS global land cover product MCD12Q1 was selected as a research object, the 1:100 000 land use product was taken as reference data, and the Beijing Tianjin Hebei region was taken as the research area to analyze and improve the target product. Firstly, linear regression models between proportion of land cover area and terrain factors were established to improve accuracy of land cover classification. The overall accuracy and Kappa coefficient of the improved product has increased by 10.96% and 0.12 compared with the original data MCD12Q1, of which the overall accuracy has increased by 0.22% in plain part and by 22.15% in non plain part. Secondly, a time-series NDVI atlas database was established, and the minimum distance method for curve similarity measurement was used to improve the accuracy of land cover classification. The overall accuracy and Kappa coefficient of the improved product has increased by 18.47% and 0.26.Thirdly,the two methods were integrated to improve the accuracy of MCD12Q1 land cover classification. Specifically, two schemes were adopted. The overall accuracy of the two schemes has improved by 11.84% and 26.28% respectively, and the Kappa coefficient has improved by 0.14 and 0.36. The conclusions are as follows: ①both terrain factors and time-series NDVI can effectively improve accuracy of land cover classification, and improvement effect based on time-series NDVI is better than that based on terrain factors. ②The improvement effect based on terrain factors is better in non plain part than that in plain part. ③Integration of terrain factors and time-series NDVI can further improve accuracy of land cover classification.
Get Citation
Copy Citation Text
Yuexin CHEN, Shunbao LIAO, Yanping WANG, Feng LI. Methods of Improving Land Cover Classification based on Terrain Factors and Time-series NDVI[J]. Remote Sensing Technology and Application, 2025, 40(2): 461
Category:
Received: May. 16, 2022
Accepted: --
Published Online: May. 23, 2025
The Author Email: LIAO Shunbao (liaoshunbao@cidp.edu.cn)