five

Size Matters: Innovative Length Estimates (SMILE) project: Length estimation of 12 target species of ecological and economical importance in the tropical western Atlantic basin, the upper Florida Keys, from 2023 to present

收藏
DataONE2025-12-10 更新2025-12-20 收录
下载链接:
https://search.dataone.org/view/10.24431_rw1k9hz_20251210T232948Z
下载链接
链接失效反馈
官方服务:
资源简介:
This package includes data file descriptions for the SMILE ("Size Matters: Innovative Length Estimates") project, that generates lengths of 12 target species in situ using a laser-mounted camera system ("FishSenseLite" or "FSL") to capture images of ecologically and economically important reef fish. The images were processed through a machine learning (AI) workflow to estimate length. To collect these images, recreational divers simply point the laser (green or red in color) at the lateral center of the fish and capture images in burst mode from a distance up to 5 meters. Through post-processing in an AI workflow (following trained datasets) using a depth of field approach, the laser location in the photo is located to identify laser position, thus distance between the camera and the subject. The position of the head and tail of the focal fish in the image is located to compute the length of the fish, given the location relevant to the distance from the camera. These data are validated against existing stereovideo methods used in concert with the FSL cameras. Thus, the data within represent lengths obtained through both stereovideo and FSL methods from August 2023 through December 2024. Length data post December 2024 was provided solely through FSL methods, unless otherwise stated. Data files include: species, site, date, camera number, file name for tracking the FSL process, length of individual, range of photo, direction of photo, points for the FSL method, common name, number of photos, habitat, depth measurements, temperature and temperatures. Data Dictionaries are included as txt files. The initial phases of this project were funded by a NOAA Coral Reef Conservation Program grant (award # FNA22NMF4410182) to develop the camera and perform field tests with citizen scientists in the upper Florida Keys as well as on regional REEF Field survey trips, with location available through the raw dataset.

本套件包含SMILE("Size Matters: Innovative Length Estimates",即"尺寸至关重要:创新体长估算")项目的数据文件说明。该项目依托搭载激光的摄像系统(FishSenseLite,简称FSL)原位测算12种目标物种的体长,用于采集具有重要生态与经济价值的礁体鱼类图像。图像将通过机器学习(AI)工作流进行处理以完成体长估算。为采集此类图像,休闲潜水员仅需将激光(颜色为绿色或红色)对准鱼类的体侧中心,并在最远5米的距离下以连拍模式拍摄图像。通过基于训练数据集的AI工作流后处理,并采用景深方法,可定位照片中的激光位置,进而确定相机与拍摄对象之间的距离。通过定位图像中目标鱼类的头部与尾部位置,并结合其与相机的距离,即可计算出该鱼类的体长。本数据集已通过与FSL摄像系统配套的现有立体视频方法进行验证。因此,本套件内的数据涵盖2023年8月至2024年12月期间通过立体视频与FSL两种方法获取的体长数据。除非另有说明,2024年12月之后的体长数据仅通过FSL方法获取。数据文件包含以下字段:物种、采样位点、采集日期、相机编号、用于追踪FSL处理流程的文件名、个体体长、照片拍摄范围、照片拍摄朝向、FSL方法所用标记点、通用名称、照片总数、栖息生境、深度测量值、温度。数据字典以纯文本(.txt)文件形式提供。本项目初期阶段由美国国家海洋和大气管理局(NOAA)珊瑚礁保护计划提供资助(资助编号:FNA22NMF4410182),用于研发该摄像系统,并与佛罗里达群岛北部的公民科学家以及区域REEF野外调查团队协作开展野外测试,原始数据集包含具体采样点位信息。
创建时间:
2025-12-10
二维码
社区交流群
二维码
科研交流群
商业服务