图像识别算法在自然资源执法中的应用研究与实践——以南京市为例

Research and Practice on the Application of Image Recognition Algorithm in Natural Resource Law Enforcement——Taking Nanjing City as an Example

为深入贯彻落实党中央、国务院关于严格耕地保护的要求,综合运用塔基监控视频、人工智能、大数据分析等技术手段强化监管,为自然资源主管部门履行“早发现、早制止、严查处”的职责提供强有力的抓手。基于塔基监控视频图像对特定要素自动识别的监管模式是实现“早发现”的主要手段,本文在既定的疑似违法建设行为分类的基础上,使用大量样本图片对开源人工智能图像识别算法进行反复训练,将训练后的算法应用于南京市自然资源执法日常工作,取得了良好的效果,构建了具有南京特色的“早发现、早识别、早预警”工作机制。

To thoroughly implement the requirements of the Party Central Committee and the State Council on strictly protecting arable land,we comprehensively apply technological means such as tower-based monitoring videos,artificial intelligence,and big data analysis to strengthen supervision,providing a powerful tool for the natural resources management department to fulfill its responsibilities of"early detection,early prevention,and strict investigation and punishment".The regulatory mode based on tower-based monitoring video image recognition for specific elements is the main means of achieving"early detection".Based on the established classification of suspected illegal construction behavior,this article uses a large number of sample images to repeatedly train open-source artificial intelligence image recognition algorithms.The trained algorithm is applied to the daily work of natural resources law enforcement in Nanjing,achieving good results and establishing a"early detection,early recognition,and early warning"mechanism with Nanjing characteristics.