A Construction Waste Landfill Dataset of Two Districts in Beijing, China from High Resolution Satellite Images

CWLD constructs and forms a construction waste landfill dataset in Changping and Daxing districts of Beijing using Gaofen-2 remote sensing satellite as the data source. The dataset contains samples of the original image area and provides mask labeled images in the semantic segmentation domain.Each pixel inside a construction waste landfill is categorized in detail according to the image background area, the open space area, the engineering facility area and the waste dumping area. It contains 237,115,531 pixels of construction waste and 49,724,513 pixels of engineering facilities. The dataset consists of three folders: Original Dataset, Construction Waste Landfill Dataset, and Deep Learning Datasets. Original Dataset. Remote sensing images and labelled images are stored separately according to Changping and Daxing districts in the Original Dataset folder. The Construction Waste Landfill Dataset folder contains the raw data enriched with data enhancement techniques. Deep Learning Datasets. The input layer of the neural network model usually needs to have a fixed input size, so it is necessary to preprocess the data before training by adjusting the input data to 512×512px, which is divided into the training set and the validation set in accordance with 8:2. Visit the GitHub page for scripts and instructions on how to use this dataset for visualizing and plotting basic statistics. The models and the code to execute them are released on https://github.com/huangleinxidimejd/CWLD_Model.

搜集汇总

背景与挑战

背景概述

该数据集利用高分二号卫星影像，构建了北京昌平和大兴两区的建筑垃圾填埋场数据集，提供语义分割标注，将像素分为背景、空地、工程设施和废弃物倾倒区四类，包含约2.37亿个建筑垃圾像素和0.5亿个工程设施像素。适用于遥感图像语义分割和建筑垃圾监测相关研究。