虎豹国家公园野生动物红外立体相机数据集（2019.10-2020.10）

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 consists of wildlife infrared stereo camera data deployed in the Hunchun Tiger and Leopard Park, Yanbian Korean Autonomous Prefecture, Jilin Province. A total of 9 wildlife infrared stereo camera devices were deployed near Erdaogou in the Tiger and Leopard Park, with each camera mounted on a tree and powered by storage batteries. The data collection period spans from October 2019 to October 2020, with a total of 5.02 GB of image data. Most of the data is stored on the SD cards inside the cameras, while some was transmitted via a 700M network. Due to signal shortages, some data remains stored in the cameras' memory cards and has not yet been retrieved. Notably, some image files have a start timestamp of January 1, 2000.

The stereo cameras capture wildlife images via infrared induction triggering. The cameras remain in a sleep state for long periods, during which only the infrared sensor is active. When the sensor detects infrared thermal information, it wakes the stereo camera to capture images. Each captured image has a resolution of 2592×1944.

1. Wildlife appearing at the installation sites in Hunchun Tiger and Leopard Park: The camera triggers a single capture whenever a wildlife animal enters the detection range of the stereo camera.
2. Data Source: Project "Development of Terrestrial Vertebrate Monitoring Equipment" (Grant No. 2016YFC0500104), completed by the Institute of Semiconductors, Chinese Academy of Sciences. The data is raw and unprocessed.
3. Photo data structure: Each valid dataset consists of a pair of left and right images. After stereo rectification, a disparity map can be generated from the paired images. Based on the disparity map, the size and distance information of target objects of interest in the images can be obtained, enabling long-term analytical research on the size information of captured animals.
4. Application scenarios: This dataset can be used to record the population size and body size of wildlife in a given area, establish a real-world wildlife body size database, obtain body size data of animals across different regions, age groups and genders, and provide supporting data for wildlife research.

AI搜集汇总

背景与挑战

背景概述

该数据集是一个用于野生动物监测的红外立体相机数据集，覆盖2019年10月至2020年10月，在吉林省虎豹国家公园部署9台相机，采集了5.01 GB的图像数据。其特点是利用红外感应触发拍摄，提供左右立体图像，可用于分析野生动物的种群数量、体型尺寸和距离信息，支持生态研究和数据库建立。