停车场车位数据集(PKLot)

这个数据库包含12,417张图像(1280X720)，这些图像是在晴朗、多云和下雨的天气下从两个不同的停车场(停车场1和停车场2)拍摄的。第一个停车场有两个不同的拍摄角度(停车场1a和停车场1b)。 这些图像被组织成三个目录(parkingla、parkinglb和parking2)。每个目录包含三个子目录，分别用于不同的天气条件(多云、下雨和晴天)。在每个子目录内，图像按获取日期组织。 数据库中的每个图像都有一个关联的XML文件，包括所有停车位的坐标及其标签(占用/空闲)通过使用XML文件对停车位进行分段，您将能够获得大约695,900张停车位的图像。This dataset contains 12,417 images with a resolution of 1280×720 pixels, captured under sunny, cloudy, and rainy weather conditions from two distinct parking lots (Parking Lot 1 and Parking Lot 2). The first parking lot has two different shooting angles (Parking Lot 1a and Parking Lot 1b). The images are organized into three directories: parkingla, parkinglb, and parking2. Each directory contains three subdirectories corresponding to different weather conditions: cloudy, rainy, and sunny. Within each subdirectory, images are sorted by their capture dates. Each image in the dataset is associated with an XML file that includes the coordinates of all parking spaces and their labels (occupied/empty). By cropping parking spaces using these XML files, approximately 695,900 parking space images can be obtained.

搜集汇总

背景与挑战

背景概述

停车场车位数据集（PKLot）是一个用于车位检测的公开数据集，包含12417张高分辨率图像，覆盖两个停车场在晴天、多云和雨天三种天气条件下的场景，每张图像均附有XML文件提供车位坐标和占用状态标注，可用于机器学习模型训练和车位识别研究。