人员离岗检测

识别人工柜台、室内工位、工厂较固定岗位的工作人员是否在岗，支持配置在岗人数。要人员活动区域较固定，且无关人员不会进入到此区域的场景。
精度对应场景或者限制条件
视频分辨率：不低于720P，码率≥1Mbps，建议2Mbps
环境要求：室内/室外，亮度＞100lux，目标清晰，画面中光线均匀，无强背光、无强逆光，无过亮过暗现象
检测目标需求：最小人体像素64*32(1080P)，人体至少1/2未被遮挡，单图人体数量小于30； 最小人头像素32*32，至少1/2未被遮挡。
色彩要求：彩色

This dataset is designed to detect whether staff working at fixed posts such as manual service counters, indoor workstations and factory fixed positions are on duty, and supports configuring the number of on-duty personnel. It is applicable to scenarios where the staff's activity area is relatively fixed and unauthorized personnel are prohibited from entering the area.

The detection accuracy varies based on the specific scenario and constraint conditions, with the following specifications:
1. Video parameters: Resolution no lower than 720P; bit rate ≥ 1 Mbps, 2 Mbps is recommended.
2. Environmental requirements: Applicable to indoor or outdoor scenarios; illumination > 100 lux; clear target subjects; uniform lighting in the frame, free of harsh backlighting, strong contre-jour lighting, over-exposed or under-exposed areas.
3. Detection target requirements: Minimum human body pixel size: 64×32 (for 1080P resolution); at least 50% of the human body must be unoccluded; the number of human bodies in a single frame shall be less than 30. Minimum human head pixel size: 32×32, with at least 50% unoccluded.
4. Color requirement: Full-color imaging.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于检测固定岗位工作人员是否在岗的算法模型，适用于人员活动区域固定且无关人员不进入的场景，对视频分辨率、环境亮度、检测目标等有具体要求。