深度学习支持下的宅基地复垦项目真实性智能审查技术研究与应用

全域综合整治背景下的宅基地复垦项目审查是保证复垦项目合法性合规性的有效手段，针对当前宅基地复垦项目审查效率低的问题，本文提出了基于遥感影像的农村宅基地复垦项目立项真实性审查方法。首先以重庆市农村宅基地复垦项目为例，建立宅基地复垦项目立项真实性审查模型，利用改进的U-Net模型准确提取建筑物，然后与复垦红线叠加实现复垦项目真实性的智能审查。试验结果表明，该方法能够快速准确地审查复垦项目真实性，对优化土地结构、促进土地流转、实现乡村振兴具有重要意义，可为全域综合整治背景下宅基地复垦项目真实性审查提供技术参考。

Research and application of intelligent verification technology for authenticity of homestead reclamation project based on deep learning

The review of homestead reclamation project under the background of region-wide comprehensive renovation is an effective means to ensure the legality and compliance of the reclamation project. Aiming at the low review efficiency of the current homestead reclamation project, a authenticity review method of rural homestead reclamation project was proposed based on remote sensing images. Taking the rural housing land reclamation project in Chongqing as an example, the paper establishes the authenticity review model of housing land reclamation project, and uses the improved U-Net to accurately extract buildings, and then superimpose with the reclamation red line to realize the intelligent authenticity review of the housing land reclamation project. The experimental results show that this method can quickly and accurately examine the authenticity of residential land reclamation projects, which is of great significance to optimize land structure, promote land circulation and promote rural revitalization. It can provide technical reference for the authenticity examination of residential land reclamation projects under the background of regional comprehensive renovation.