德清水域数字化遥感监测识别数据

用于实现对水域监测中问题点位的高效精准识别，包括水面垃圾漂浮、水体富营养化、破坏河道或损毁、山塘水库变化等现象的识别。能够监测出水域环境的异常，及时发现水域变化区域。自动识别水域变化区域坐标，便于确定需要进行实地处理的位置。自动识别监测内容和变化类型，便于掌握实地情况。为后续的问题派发提供方便，有利于工作人员及时掌握水域异常情况，避免河湖等地出现问题无人管理而引发自然灾害。通过遥感监测解决了网格员在传统人工巡查中难以发现问题，容易忽略问题，巡查范围太大，隐蔽地段、偏远地区与恶劣环境中不易巡查等问题。将系统识别出的问题点位派发给网格员，使网格员有依据、有目标的实地确认，极大的提高了发现问题的效率，节约人员时间与人工成本，避免网格员出现无效的巡查。基于无人机航拍采集真实水域环境遥感影像数据，通过YOLO算法进行实时目标检测。首先将单元神经网络应用于完整的遥感影像，将图像分割成19x19的单元格，每个单元神经网络负责预测K个单元格。预测每个区域的概率，所有单元格上具有最大概率的类被选择并分配给特定的网格单元，生成由预测概率加权的边界框。
在预测类概率后，进行NMS运算，算法识别下一个最高类别概率的边界框，并进行相同的运算过程，直到剩下所有不同的边界框。算法输出所需的向量，并显示各个类的边界框的细节。
通过判断结果正确或错误来纳入或排除数据，识别出水面垃圾漂浮、水体富营养化、河道占用或损毁、山塘水库变化等现象，自动获取问题点位坐标，匹配到村镇级行政区划，并将影像中的特征区域与坐标等信息自动上传至德清水利数字化遥感监测平台，最终获得德清水利数字化遥感监测识别数据。

This dataset is developed for efficient and accurate identification of problematic spots during water area monitoring, covering phenomena such as floating surface waste, water eutrophication, damaged or occupied river channels, and changes in ponds and reservoirs. It can detect water environment abnormalities and timely identify areas with water body changes. It automatically recognizes the coordinates of water change areas, enabling the confirmation of locations requiring on-site treatment. It also automatically identifies monitoring contents and change types, which helps to grasp on-site situations. This facilitates subsequent task dispatching, allows staff to timely grasp water abnormalities, and prevents natural disasters caused by unmanaged issues in rivers and lakes.

The remote sensing monitoring-based solution addresses the pain points of grid inspectors in traditional manual patrols, including difficulty in detecting problems, proneness to overlooking issues, overly large patrol scope, and inconvenient patrols in hidden areas, remote regions and harsh environments. Dispatching the identified problematic spots to grid inspectors enables them to conduct on-site verification with clear targets and basis, greatly improving problem detection efficiency, saving labor time and costs, and eliminating ineffective patrols by inspectors.

The system collects real-world remote sensing image data of water environments via unmanned aerial vehicle (UAV) aerial photography, and performs real-time object detection through the YOLO algorithm. First, the cell neural network is applied to the complete remote sensing image, which is split into 19×19 grid cells. Each cell neural network is responsible for predicting K bounding boxes within its assigned cell. The confidence probability of each region is predicted, and the class with the maximum probability across all predictions is selected and assigned to the specific grid cell, generating bounding boxes weighted by the prediction probabilities.

After class probability prediction, the Non-Maximum Suppression (NMS) operation is conducted. The algorithm identifies the bounding box with the next highest class probability and repeats the same procedure until all distinct bounding boxes are retained. The algorithm outputs the required vectors and displays detailed information of the bounding boxes for each category.

Data entries are included or excluded based on the correctness of the recognition results. The system identifies target phenomena including floating surface waste, water eutrophication, occupied or damaged river channels, and changes in ponds and reservoirs, automatically acquires the coordinates of problematic spots, matches them to village and town-level administrative divisions, and automatically uploads information such as feature regions and coordinates in the images to the Deqing Water Conservancy Digital Remote Sensing Monitoring Platform, ultimately generating the Deqing Water Conservancy Digital Remote Sensing Monitoring and Recognition Dataset.

搜集汇总

数据集介绍

特点

德清水域数字化遥感监测识别数据是由浙江国遥地理信息技术有限公司产生的数据集，包含500条记录，用于高效精准识别水域监测中的问题点位。数据通过无人机航拍和YOLO算法实时检测，识别水面垃圾、水体富营养化等问题，并上传至监测平台。