基于多源时序影像的水稻分布提取及变化分析

Rice Distribution Extraction and Change Analysis Based on Multi-Source Time-Series Images

摘要

高效精确地提取水稻分布范围对农作物产量分析及调整农业规划管理起到关键作用。该研究基于Google Earth Engine（GEE）云平台利用多时相Sentinel影像等多源数据，通过多种机器学习算法对2016年盘锦市进行监督分类，通过连续变化检测与分类（Continuous Change Detection and Classification，CCDC）算法更新2017-2024年间水稻时空分布情况，进一步探讨水稻在不同年份间的动态变化。结果表明： （1）综合使用遥感-地形特征构建水稻提取模型及分时段的特征优选方法均有效地提升了分类模型的性能和准确性。 （2）基于瓦片的随机森林分类方法（TB_RF）在水稻提取过程中展现出较高的精度和适用性。其获取的 2016 年盘锦市水稻分布范围精度最高，总体精度（Overall Accuracy，OA）和Kappa系数分别为95.4%、0.91，水稻种植面积与实际面积误差小于1%。 （3）CCDC算法在利用Sentinel影像进行盘锦市水稻长时序提取研究中表现良好。CCDC算法在2021年的分类结果精度最低，其OA和Kappa系数分别为95%、0.90。历年的水稻种植面积的计算值与真实面积之间的差值均保持在0.5%以下。 （4）2016-2024年盘锦市水稻分布范围没有明显变化，分布面积缓慢增长，呈现出“总体稳定、局部调整、缓慢扩张”的演变趋势。本研究结果可为水稻种植发展规划与监测分析提供科学参考依据。

Abstract

Efficient and accurate rice distribution extraction is crucial for analyzing crop yields and adjusting agricultural planning and management. Based on Google Earth Engine(GEE) cloud platform, this study uses multi-source data including multi-temporal Sentinel images,to supervise and classify Panjin City in 2016 through various machine learning algorithms, and updates the spatial and temporal distribution of rice from 2017 to 2024 through CCDC algorithm. Additionally, it further explores the dynamic changes of rice in different years. The results show that: (1) The comprehensive use of remote sensing- terrain features to build a rice extraction model and the time-division feature optimization method effectively improves the performance and accuracy of the classification model. (2) The tile-based random forest classification method (TB_RF) demonstrates high accuracy and applicability in rice extraction. The extracted distribution range of rice in Panjin City in 2016 achieves the highest precision, with an Overall Accuracy (OA) of 95.4% and a Kappa coefficient of 0.91. The error between the calculated rice planting area and the actual area is less than 1%. (3) CCDC algorithm performs well in the long- time series extraction in Panjin City using Sentinel images. The classification results for 2021 have the lowest accuracy, with an OA of 95% and a Kappa coefficient of 0.90. The difference between the calculated rice planting area and the actual area across the years remains below 0.5%. (4) From 2016 to 2024, the distribution range of rice in Panjin City shows no significant changes, with a slow expansion in area, reflecting an evolution trend characterized by "overall stability, local adjustment and gradual expansion". The findings of this study can provide a scientific reference for the planning, monitoring, and analysis of rice cultivation development.

关键词

遥感 / 水稻 / 分类 / Google Earth Engine / CCDC / Sentinel-1/2 / 时序分析

Key words

Remote sensing / rice / classification / Google Earth Engine / CCDC / Sentinel-1/2 / time series analysis

参考文献

[1] 丁文家, 胡峻铭, 王嘉力. 水稻育种主要目标性状基因挖掘研究进展[J]. 杂交水稻, 2023, 38(03): 1-19.

[2] 张合红, 李雁军, 孙宗涛, 等. 水稻病毒与植物寄主互作研究新进展[J]. 植物保护学报, 2022, 49(01): 33-41.

[3] 张德咏. 袁隆平与杂交水稻科技创新[J]. 杂交水稻, 2022, 37(S1): 234-235.

[4] 谭 深, 吴炳方, 张 鑫. 基于Google Earth Engine 与多源遥感数据的海南水稻分类研究[J]. 地球信息科学学报, 2019, 21(06): 937-947.

[5] 孙亚楠, 李仙岳, 史海滨, 等. 基于特征优选决策树模型的河套灌区土地利用分类[J]. 农业工程学报, 2021, 37(13): 242-251.

[6] 张天一,苏 华,杨 欣, 等. 基于LightGBM 的全球海洋次表层温盐遥感预测[J]. 遥感学报, 2020, 24(10): 1255-1269.

[7] 张 颖, 何贞铭, 吴贞江. 基于多源遥感影像的农作物分类提取[J]. 山东农业大学学报(自然科学版), 2021, 52(04): 615-618.

[8] 李晓慧,王 宏, 李晓兵, 等 . 基于多时相 Landsat 8 OLI影像的农作物遥感分类研究[J]. 遥感技术与应用, 2019, 34(02): 389-397.

[9] Du L, Li Z, Wang Q, et al. An optimized semi-supervised generative adversarial network rice extraction method based on time-series sentinel images[J]. Agriculture, 2024, 14(9): 1505.

[10] 朱 磊,王 科, 丁一民, 等. 基于Sentinel-2 的青铜峡灌区水稻和玉米种植分布早期识别[J]. 干旱区地理, 2024, 47(05): 850-860.

[11] 周兴霞, 王颖洁, 杨 攀. 基于光学与雷达遥感影像协作的多云雾区域农作物信息提取研究[J]. 遥感技术与应用, 2024, 39(02): 362-372.

[12] Pan B, Zheng Y, Shen R, et al. High resolution distribution dataset of double-season paddy rice in China[J]. Remote Sensing, 2021, 13(22): 4609.

[13] Fiorillo E, Di Giuseppe E, Fontanelli G, et al. Lowland rice mapping in Sédhiou Region (Senegal) using sentinel 1 and sentinel 2 data and random forest[J]. Remote Sensing, 2020, 12(20): 3403.

[14] 乐 颖, 刘聚涛, 文 慧. 基于多源多时相影像的鄱阳湖水体提取及时空变化分析[J]. 长江科学院院报, 2024, 41(08): 164-171+ 206.

[15] 李恒凯, 贺明华, 王秀丽. 时序SAR与光学影像的南方丘陵区水稻种植识别[J]. 测绘科学, 2023, 48(05): 78-86.

[16] He H, Yan J, Liang D, et al. Time-series land cover change detection using deep learning-based temporal semantic segmentation[J]. Remote Sensing of Environment, 2024, 305: 114101.

[17] Zhu Z, Woodcock C E. Continuous change detection and classification of land cover using all available Landsat data[J]. Remote Sensing of Environment, 2014, 144: 152-171.

[18] Xie S, Liu L, Zhang X, et al. Mapping the annual dynamics of land cover in Beijing from 2001 to 2020 using Landsat dense time series stack[J]. Isprs Journal of Photogrammetry and Remote Sensing, 2022, 185: 201-218.

[19] Vogelmann J E, Gallant A L, Shi H, et al. Perspectives on monitoring gradual change across the continuity of Landsat sensors using time-series data[J]. Remote Sensing of Environment, 2016, 185: 258-270.

[20] Fu P, Weng Q. A time series analysis of urbanization induced land use and land cover change and its impact on land surface temperature with Landsat imagery[J]. Remote Sensing of Environment, 2016, 175: 205-214.

[21] 王春玲, 史锴源,庞 勇, 等. 基于变化更新的中东欧国家森林覆盖制图[J]. 林业科学, 2024, 60(05): 116-126.

[22] 苏 岑. 基于CCDC 和LandTrend 算法的衡邵干旱走廊气象干旱时空特征研究[J]. 国土资源导刊, 2024, 21(02): 158-164.

[23] 肖 真, 丁明军, 刘怡媛, 等. Google Earth Engine平台下的1990-2019 年江西省森林覆被连续变化分析[J]. 测绘通报, 2022(03): 18-22.

[24] Jiang N, Li P, Feng Z. Detecting tropical freshly-opened swidden fields using a combined algorithm of continuous change detection and support vector machine[J]. International Journal of Applied Earth Observation and Geoinformation, 2025, 136: 104403.

[25] Demarquet Q, Rapinel S, Gore O, et al. Continuous change detection outperforms traditional post-classification change detection for long-term monitoring of wetlands[J]. International Journal of Applied Earth Observation and Geoinformation, 2024, 133: 104142.

[26] 任鸿瑞, 张悦琦, 何奇瑾, 等. 基于FY-3_MERSI遥感数据的水稻种植分布提取[J]. 光谱学与光谱分析, 2023, 43(05): 1606-1611.

[27] 贾永倩, 王振锡, 瞿余红, 等. 基于全极化Radarsat-2数据的苹果园平均树高估算模型[J]. 森林工程, 2017, 33(02): 45-49+ 55.

[28] Breiman L. Random Forests[J]. Machine Learning, 2001, 45(1): 5-32.

[29] Cortes C, Vapnik V. Support-Vector Networks[J]. Machine Learning, 1995, 20(3): 273-297.

[30] Li B, Friedman J, Olshen R, et al. Classification and Regression Trees (CART)[J]. Biometrics, 1984, 40(3): 358-361.

[31] Zhang M, Huang H, Li Z, et al. Automatic high-resolution land cover production in madagascar using Sentinel-2 time series, tile-based image classification and google earth engine[J]. Remote Sensing, 2020, 12(21): 3663.

基金资助

中国气象局人才项目“中国气象局农业气象重点创新团队”(CMA2024ZD02)

科院基本科研业务费(2024Z001)