虎豹国家公园野生动物红外立体相机数据集（2019.10-2020.10）
收藏资源简介：
本数据集包含部署在吉林省延边朝鲜族自治州珲春市虎豹公园的野生动物红外立体相机的数据集。总计有9台野生动物红外立体相机设备部署在虎豹公园二道沟附近，立体相机分别置于树上，采用蓄电池进行供电。本数据采集时间为2019年10月至2020年10月内，共计5.02G图像数据，数据存储在相机内部SD卡内，部分通过700M网络进行传输，因无信号，部分数据保存在相机内部存储卡中暂未取出，相机部分图像起始时间为2000年1月1日。 立体相机采用红外感应触发获取野生动物图像，相机长期处于睡眠状态，此状态下仅红外传感器处于工作状态，当传感器感应到红外热信息唤起立体相机进行拍照采集，采集获得图像单张大小为2592×1944。 1. 珲春虎豹公园安装点出没的野生动物。有野生动物出现在立体相机探测范围内时，即触发采集拍照一次。 2. 数据来源："陆生脊椎动物监测设备研制“课题，2016YFC0500104，完成单位：中国科学院半导体研究所，原始数据，未加工。 3. 照片数据分为一对有效数据包括左图和右图，校正后左右图像可获得视差图，根据视差图可获取图像中感兴趣目标的尺寸信息和距离信息，通过对获得的动物尺寸信息进行长期分析研究。 4. 本数据可用于记录一定区域内的野生动物种群数量和体型大小，可建立真实野生动物体型数据库，获取不同区域、年龄和性别下的动物体型数据信息，为野生动物研究提供支撑数据。
This dataset contains data collected from wildlife infrared stereo camera traps deployed in the Hunchun Tiger and Leopard Park, Yanbian Korean Autonomous Prefecture, Jilin Province. A total of 9 such camera traps were deployed near Erdaogou in the Tiger and Leopard Park, mounted on trees and powered by storage batteries. The data collection period spanned from October 2019 to October 2020, with a total of 5.02 GB of image data. Most of the data was stored on the SD memory cards inside the cameras, and part of it was transmitted via a 700M network. However, due to lack of network signal, some data remained stored in the cameras' memory cards and had not been retrieved yet. Notably, the starting capture timestamp of some images from the cameras is January 1, 2000. These stereo camera traps capture wildlife images via infrared sensing triggering. The cameras remain in a low-power sleep mode for long periods, during which only the infrared sensor is active. When the sensor detects infrared thermal information, it wakes up the stereo camera to perform image capture. Each captured image has a resolution of 2592 × 1944 pixels. 1. Wildlife recorded at the deployment sites in Hunchun Tiger and Leopard Park: The camera trap will trigger one capture whenever a wild animal enters its detection range. 2. Data Source: Funded by the project "Development of Terrestrial Vertebrate Monitoring Equipment" (Grant No. 2016YFC0500104), completed by the Institute of Semiconductors, Chinese Academy of Sciences. This is raw, unprocessed original data. 3. The photo data consists of valid stereo pairs including left and right images. After calibration, a disparity map can be generated from the left and right images. The disparity map allows extraction of size and distance information of targets of interest in the images, enabling long-term analytical studies based on the acquired animal size data. 4. This dataset can be used to record wildlife population size and body size within a specific area. It can support the establishment of a real-world wildlife body size database, collecting animal body size data across different regions, age groups and genders, providing supporting data for wildlife research.
