火焰检测

针对室内（办公室、卧室、仓库、楼道等）场景下对火焰进行快速、精准的检测识别，对防火消防起到辅助作用，实现有效的城市监管。
精度对应场景或者限制条件
最小尺寸：100*100px（1080p）
遮挡：遮挡<50%
输入源：RGB图片
场景：夜间或白天

This dataset is designed for fast and accurate flame detection and recognition in indoor scenarios such as offices, bedrooms, warehouses, corridors and other similar indoor environments, to assist fire prevention and firefighting efforts and enable effective urban supervision.

Relevant constraints and applicable scenarios for detection accuracy are as follows:
1. Minimum target size: 100×100 pixels (under 1080p resolution)
2. Occlusion limit: allowable occlusion rate < 50%
3. Input source: RGB images
4. Applicable scenarios: daytime or nighttime

搜集汇总

背景与挑战

背景概述

火焰检测数据集针对室内场景（如办公室、卧室、仓库、楼道等）设计，用于快速、精准地检测火焰，辅助防火消防和城市监管。数据集要求最小尺寸为100*100px（1080p），遮挡小于50%，输入源为RGB图片，适用于夜间或白天场景。