及时准确获取土地利用空间格局演变规律,能够有效为城市生态环境保护和科学管理提供依据。文章利用卷积神经网络提取多个时期土地利用空间特征,结合多种空间驱动因子建立长短时记忆网络(long short term memory network,LSTM)的元胞自动机(cellular automata,CA)模型(LSTM-CA)。以张家口市中心城区1995年、2000年、2005年、2010年及2015年5期时序土地利用分类、地形及城市交通等数据为基础,开展2020年城市土地利用模拟方法研究。通过与多层感知机(multi-layer perceptron,MLP)的CA模型(MLP-CA)进行精度对比分析,结果表明所提方法Kappa系数达到0.90,FoM指标为0.39,各项指标均优于MLP-CA模型, LSTM-CA更能充分挖掘历史土地利用变化之间的内在关系,可以有效提升模拟精度。
Timely and accurate acquisition of the spatial pattern evolution of land use can effectively support urban ecological environment protection and scientific management. In this study, the spatial characteristics of land use in multiple periods were extracted using a convolutional neural network. Then, they were combined with multiple spatial driving factors to build the long short-term memory network - cellular automata (LSTM-CA) model. Based on the data of land use classification, terrain, and urban traffic of the Zhangjiakou central urban area in 1995, 2000, 2005, 2010, and 2015, this study investigated the simulation methods for urban land use in 2020. By comparison with the precision of the multi-layer perceptron - cellular automata (MLP-CA) model, the proposed method has a Kappa coefficient of over 0.90 and FoM of 0.39. All indices are better than those of the MLP-CA model. Therefore, the LSTM-CA model can fully explore the internal relationships between the changes in historical land use and effectively improve the simulation precision.
刘春霖, 夏建新. 基于LSTM-CA模型的土地利用动态模拟[J]. 自然资源遥感, 2022, 34(4): 122-128.
LIU Chunlin, XIA Jianxin. Dynamic simulation of land use based on the LSTM-CA model. Remote Sensing for Natural Resources, 2022, 34(4): 122-128.
Liao J F, Tang L N, Wang C P, et al. Measuring and calibrating extended neighborhood effect of urban cellular automata model based on particle swam optimization[J]. Progress in Geography, 2014, 33(12):1624-1633.
Chen B F, Zhang Y M, Jiang D. Urban land expansion in Fuzhou City based on coupled cellular automata and agent-based models (CA-ABM)[J]. Progress in Geography, 2017, 36(5):626-634.
doi: 10.18306/dlkxjz.2017.05.010
Sun Y Z, Yang J, Song S Y, et al. Modeling of multi-level vector cellular automata and its simulation of land use change based on urban planning[J]. Acta Geographica Sinica, 2020, 75(10):2164-2179.
Tian J M, Chen J. Simulation and prediction of land use in the high standrad grain area of Hebi City[J]. Remote Sensing for Land and Resources, 2018, 30(1):150-156.doi:10.6046/gtzyyg.2018.01.21.
doi: 10.6046/gtzyyg.2018.01.21
Yang J, Xie P, Xi J C, et al. LUCC simulation based on the cellular automata simulation:A case study of Dalian Economic and Technological Development Zone[J]. Acta Geographica Sinica, 2015, 70(3):461-475.
Zhang D C, Liu X P, Yao Y, et al. Simulating spatiotemporal change of multiple land use types in Dongguan by using random forest based on cellular automata[J]. Geography and Geo-Information Science, 2016, 32(5):29-36.
[7]
Xing W, Qian Y, Guan X, et al. A novel cellular automata model integrated with deep learning for dynamic spatio-temporal land use change simulation[J]. Computers and Geosciences, 2020, 137(1):1-9.
[8]
Von Neumann J. Theory of self-reproducing automata[J]. London:University of Illinois Press, 1966:3-14.
[9]
Karimi F, Sultana S, Babakan A S, et al. An enhanced support vector machine model for urban expansion prediction[J]. Computers,Environment and Urban Systems, 2019, 75(1):61-75.
doi: 10.1016/j.compenvurbsys.2019.01.001
[10]
Shafizadeh-Moghadam H, Asghari A, Tayyebi A, et al. Coupling machine learning,tree-based and statistical models with cellular automata to simulate urban growth[J]. Computers,Environment and Urban Systems, 2017, 64(1):297-308.
doi: 10.1016/j.compenvurbsys.2017.04.002
[11]
He J, Li X, Yao Y, et al. Mining transition rules of cellular automata for simulating urban expansion by using the deep learning techniques[J]. International Journal of Geographical Information Science, 2018, 32(10):2076-2097.
doi: 10.1080/13658816.2018.1480783
[12]
Gounaridis D, Chorianopoulos I, Symeonakis E, et al. A random forest-cellular automata modelling approach to explore future land use/cover change in Attica (Greece),under different socio-economic realities and scales[J]. Science of the Total Environment, 2019, 64(6):320-335.
[13]
Grekousis G. Artificial neural networks and deep learning in urban geography:A systematic review and meta-analysis[J]. Computers,Environment and Urban Systems, 2019, 74(1):244-256.
doi: 10.1016/j.compenvurbsys.2018.10.008
[14]
Guan D, Zhao Z, Tan J. Dynamic simulation of land use change based on logistic-CA-Markov and WLC-CA-Markov models:A case study in three gorges reservoir area of Chongqing,China[J]. Environmental Science and Pollution Research, 2019, 26(20):20669-20688.
doi: 10.1007/s11356-019-05127-9
[15]
Jia X, Khandelwal A, Nayak G, et al. Incremental dual-memory lstm in land cover prediction[C]// Proceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data mining, 2017:867-876.
[16]
Wang H, Zhao X, Zhang X, et al. Long time series land cover classification in China from 1982 to 2015 based on Bi-LSTM deep learning[J]. Remote Sensing, 2019, 11(14):1-22.
doi: 10.3390/rs11010001
[17]
White R, Engelen G. Cellular automata and fractal urban form:A cellular modelling approach to the evolution of urban land-use patterns[J]. Environment and Planning A, 1993, 25(8):1175-1199.
doi: 10.1068/a251175
[18]
Liu X, Liang X, Li X, et al. A future land use simulation model (FLUS) for simulating multiple land use scenarios by coupling human and natural effects[J]. Landscape and Urban Planning, 2017, 168(1):94-116.
doi: 10.1016/j.landurbplan.2017.09.019
2022, Vol. 34 
), 夏建新(
