Skip to content

Realize the detection of non-motorized vehicles (including two-wheeled vehicles and tricycles) and pedestrians in highway and urban night scenes (black days and fully illuminated street lights).

Notifications You must be signed in to change notification settings

Linamiao-1998/task5_vehicles_detection

Repository files navigation

项目简介:

使用YOLOV5实现高速公路和城区夜晚场景下(黑天、路灯全亮)非机动车(包括二轮车、三轮车)和行人的检测。

项目文件说明:

competition_yolov5

————models       # yolov5模型和配置文件

————utils        # yolov5检测器、评价指标和数据加载文件

————weights      # yolov5模型权重

————task5        # 待测数据

————task5_result # 待测数据测试结果,保存为json文件

————Detector.py  # 将yolov5检测器封装为一个类

————demo.py      # 调用yolov5检测器类进行检测

————results.mp4  # 待测数据测试结果视频

使用说明:

1、运行

python demo.py

2、修改参数 可在 ./utils/BaseDetector.py 中对self.threshold(置信度阈值)进行修改; 在Detector.py中第51行pred = non_max_suppression(pred, self.threshold, 0.1)对最后一个参数(nms阈值)进行修改。

最终效果:

可查看results.mp4

About

Realize the detection of non-motorized vehicles (including two-wheeled vehicles and tricycles) and pedestrians in highway and urban night scenes (black days and fully illuminated street lights).

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages