基于深度学习的耕地新增建房监测研究

Research on Monitoring of New Buildings on Cultivated Land Based on Deep Learning

基于深度学习技术,针对福建省常态化新增农村乱占耕地建房监测需求,构建了混合分辨率新增建房监测遥感影像样本数据集,采用卷积神经网络进行深度学习模型训练,并利用训练后的最优模型开展了耕地新增建房提取应用实验。结果表明,基于深度学习的耕地新增建房监测技术具备一定的应用潜力,可在实际工程中进行应用探索,以提高监测工作的效率和精度。

Based on the deep learning technology,according to the normalized demand of new buildings on cultivated land monitoring in Fujian Province,we constructed a sample dataset of multi-resolution remote sensing images of new buildings on cultivated land monitoring.We used the convolutional neural network to train deep learning model at first.And then,based on the optimal model,we carried out the application test of new buildings extraction on cultivated land.The results show that the new buildings on cultivated land monitoring technology based on deep learning has a certain application potential,which can be applied in practical engineering to improve the efficiency and precision of monitoring work.