本研究基于水稻孕穗期、抽穗期、灌浆期和成熟期4个生育期的Sentinel-2遥感数据，分析各生育期内卫星遥感光谱参数与稻米品质指标的关系，建立基于各生育期卫星光谱信息的水稻品质指标预测模型。将5种稻米品质指标分别与4个生育期内的光谱参数进行皮尔逊相关性分析，结果表明，5项品质指标在4个生育期内均与光谱参数有不同程度相关性。然后筛选出相关性效果显著的光谱参数，用于建立各品质指标的预测方程，建模结果表明，基于卫星遥感光谱信息解释率由大到小的稻米品质指标依次是精米率>长宽比>蛋白质含量>直链淀粉含量>糙米率；卫星遥感光谱反演稻米各品质指标所在的最佳生育期不同，糙米率和精米率的最佳生育期为抽穗期，其建模决定系数（Coefficient of Determination，R²）分别为0.461和0.893；长宽比的最佳生育期为成熟期，R²为0.878；直链淀粉含量和蛋白质含量的最佳生育期为灌浆期，R²分别为0.646和0.647；基于卫星遥感光谱信息的稻米品质模型验证效果较好，解释率为51%~74%。可见，利用卫星遥感技术能够实现大范围水稻品质指标定量监测与评估。

Based on Sentinel-2 remote sensing data of rice at booting stage, heading stage, filling stage and mature stage, this study analyzed the relationship between satellite remote sensing spectral parameters and rice quality indicators in each growth period, and established a prediction model of rice quality indicators based on satellite spectral information in each growth period, carried out pearson correlation between five different quality indexes of rice grains and spectral parameters in four growth stages. The results showed that, the five quality indicators had significant correlation with spectral parameters in different degrees during the four growth stages. Then, the spectral parameters with significant correlation effect were selected to establish the prediction equation of rice quality indicator, the modeling results showed that: (1) Based on the interpretation rate of satellite remote sensing spectral information, the rice quality indicators from large to small are: milled rice rate>length-width ratio>protein content>amylose content>brown rice rate. (2) The best growth period of rice quality indicator inverted by satellite remote sensing spectrum was different. The best growth period of brown rice rate and milled rice rate was heading stage, and the coefficient of determination (R²) was 0.461 and 0.893, respectively. The best growth period of length-width ratio was mature period, and R² was 0.878. The best growth period of amylose content and protein content was filling stage, and R² was 0.646 and 0.647, respectively. (3) The rice quality model based on satellite remote sensing spectral information had a good verification effect, and the interpretation rate was 51%-74%. Therefore, the use of satellite remote sensing technology can realize the quantitative monitoring and evaluation of rice quality indicators in a wide range.