fansel/nordland
收藏资源简介:
Nordland数据集(Olid等人分区版)是经过预处理和分区的Nordland数据集版本,用于研究季节性变化下的单视图地点识别。数据集来源于挪威广播公司(NRK)的视频,包含四个季节(春、夏、秋、冬)的图像,每秒钟提取一帧,过滤掉隧道和火车停靠点,共28,865张图像/季节。数据集分为训练集和测试集,训练集包含两个部分(12,837和11,733张图像),测试集包含三个部分(每部分1,150张图像)。每个样本包含图像(PNG格式)、季节、帧ID(相同帧ID表示相同地理位置)和部分信息(训练或测试部分)。
The Nordland Dataset (Olid et al. Partition) is a pre-processed and partitioned version of the Nordland Dataset, designed for single-view place recognition under seasonal changes. The dataset is derived from videos by the Norwegian Broadcasting Corporation (NRK) and includes images from four seasons (fall, spring, summer, winter), with one frame extracted per second. Tunnels and train stops were filtered out, resulting in 28,865 images per season. The dataset is split into training and test sets: the training set consists of two sections (12,837 and 11,733 images), and the test set consists of three sections (1,150 images each). Each sample contains an image (PNG format), season, frame ID (same frame ID indicates the same geographic location), and section information (train or test sections).




