Project Demo Video
The project demo video can be jumped to Bilibili to watch: https://www.bilibili.com/video/BV1RT411Z7kD/?vd_source=805c57038e291405fe38f3adefa0f2d2
Project Brief Introduction
This project uses Yolov5+DeepSort to track vehicles and pedestrians, and counts the number of targets in each category in real time, as well as measures the speed and acceleration of target movement, and marks and saves speeding vehicles.
– The project supports the detection of high-resolution videos, you can use **sliding window detection**, the specific method is to cut the picture of each frame according to the specified sliding step and window size, for example, the slice cut into the size of 512*512 is input into the model for inference, and then the inference results of all slices are merged, and a non-maximum suppression is required to remove the overlapping boxes detected by different slices.
– The pre-trained model of this project uses the yolov5s pre-trained weights provided by YOLOv5, and users can change their own model weight files.
– In this project, you can specify the categories to be detected and count the number of targets in each category in each frame in real time.
– This project can calculate the movement speed and acceleration of each target in real time.
– For speeding vehicles, they can be marked and saved for easy management by the traffic department.
For more information on how to obtain the full project, please refer to: https://www.bilibili.com/video/BV1RT411Z7kD/?vd_source=805c57038e291405fe38f3adefa0f2d2
Reference items
[https://blog.csdn.net/WhiffeYF/article/details/121346483] (https://blog.csdn.net/WhiffeYF/article/details/121346483)
[https://github.com/ultralytics/yolov5] (https://github.com/ultralytics/yolov5)
