| 摘要: |
| 【目的】耕地质量定级评价工作是准确把握耕地效益水平，有效保护耕地和科学管理
耕地的必要前提。【方法】以襄阳市城区耕地为研究对象，从自然要素、社会经济要素、区
位因素3 方面建立耕地质量评价体系，借助BP 神经网络法，通过训练建立网络模型，并以
仿真练习得出襄阳市城区耕地质量评价结果。【结果】（1）BP 神经网络模型的精度达到0.93
左右，建立的网络模型具有一定可行性和精确性；（2）襄阳市城区耕地质量总体水平良好，
其中二级地和三级地最多，面积占比分别为33.56% 和46.19%；耕地质量空间分布上呈现由
东北向西南递减的趋势。【结论】借助BP 神经网络的方法开展耕地质量评价具有一定的可行
性和精确性，可为以后的耕地质量评价提供一定借鉴与参考。 |
| 关键词: 耕地质量 BP 神经网络模型 耕地定级 襄阳市城区 |
| DOI：10.12105/j.issn.1672-0423.20190208 |
| 分类号: |
| 基金项目:国家自然科学基金项目（41401232）；华中师范大学中央高校基本科研业务费（CCNU18TS002） |
|
| Evaluation of cultivated land quality based on BP neural networkmethod——a case study at Xiangyang urban area |
|
Zhang Zhaohui, Nie Yan※, Ma Zeyue
|
|
Hubei Provincial Key Laboratory for Geographical Process Analysis and Simulation，
Central China Normal University，Wuhan 430062，China
|
| Abstract: |
| ［Purpose］The evaluation and grading of cultivated land quality is the necessary
premise for accurately grasping the level of cultivated land benefits，effectively protecting
cultivated land and scientifically managing cultivated land.［Method］Taking cultivated land
in Xiangyang urban area as the research object，the study establishes the evaluation system of
cultivated land quality from three aspects：natural factors，socio-economic factors and location
factors. BP neural network is used as the evaluation method of cultivated land quality，and the
network model is established through training. Simulation exercises yielded the results of the
evaluation of cultivated land quality in Xiangyang urban area.［Result］The results showed as
follows：firstly，The accuracy of the BP neural network model is about 0.93，and the established
network model has certain feasibility and accuracy. Secondly，the overall quality of cultivated
land in Xiangyang urban area is good，with the second and third grades being the largest，
accounting for 33.56% and 46.19% respectively；the distribution of cultivated land is less in the
urban area center and more in suburbs；the quality of cultivated land is declining from northeast
to southwest.［Conclusion］This research proved that The BP neural network method can be
applied to the evaluation of cultivated land quality，and it is feasible and accurate. It can provide
some reference for the future evaluation of cultivated land quality. |
| Key words: cultivated land quality BP neural network model cultivated land grading Xiangyang urban area |
