关键词: 模糊C均值, 黄河三角洲, 多尺度高斯匹配滤波, 潮沟提取, 高分二号, 异质背景

Abstract: The quality evaluation system based on meteorological elements (climate quality) is more convenient and acceptable than that based on physiological and biochemical indices. However, the simulation accuracy of climate quality model needs to be improved. Based on the observational datasets of peach quality and meteorology in Fenghua (2006-2016) and Cixi (2015-2016) in Zhejiang Province, the climate quality of juicy peach was evaluated. The Monte Carlo method and threat score (TS) analysis method were employed to improve the accuracy of the integrated quality evaluation. The spatial and temporal variations of peach quality in Zhejiang were analyzed using the climate quality model. The Monte Carlo method well reduced the uncertainty of integrated quality by 21% (16%-26%). The prediction accuracy of ensemble model constructed by TS score analysis method was significantly higher than that of a single model. The correlation coefficient between the integrated qualities established by the climatic quality ensemble model and biochemical factors based model was as high as 0.97, with the absolute error and root mean square error being 0.01 and 0.02, respectively. From 1971 to 2000, the climate trend rate of peach quality in Zhejiang Province was -0.02·10 a^-1. In the early 21st century, the climate trend rate reached -0.05·10 a^-1, and the annual fluctuation increased. The climate quality of peaches in Zhejiang Province was 0.55 (0.49-0.63), decreasing from coastal regions to inland. The climatic quality evaluation model based on Monte Carlo method and TS score analysis can well simulate the integrated quality of juicy peach in Zhejiang Province, providing an effective method for the refined assessment of peach quality at regional scale.

Key words: heterogeneous background, tidal creeks extraction, GF-2, Yellow River Delta, fuzzy C-means algorithm., multi-scale Gaussian-matched filtering