# CarDetection

**Repository Path**: evianchu/car-detection

## Basic Information

- **Project Name**: CarDetection
- **Description**: An open-source project focused on car detection, providing robust tools and algorithms for identifying vehicles in images and videos, suitable for traffic monitoring and autonomous driving systems.
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-11-25
- **Last Updated**: 2025-11-27

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Model Training and Evaluation for Car Detection
## 1. Environment Setup
``` shell
mamba create -n CarDetection
mamba activate CarDetection
pip install -r requirements.txt
```

## 2. Data Preparation
### 2.1 Prepare BIT-Vehicle Dataset
    - Download the BIT-Vehicle dataset from the [site](https://pan.baidu.com/s/1KbmQ3B7uK29d8O4Z5Cxxwg).
    - Extract the dataset to the `./data/bit vehicle/dataset` directory.
    - Extract the annotations to the `./data/bit vehicle/annotations` directory.
    - Ensure the directory structure is as follows:
      ./data/bit vehicle/
      ├── dataset/
      │   ├── vehicle_0000001.jpg
      │   ├── vehicle_0000002.jpg
      │   └── ...
      └── annotations/
          ├── vehicle_0000001.xml
          ├── vehicle_0000002.xml
          └── ...
### 2.2 Dataset Splitting
``` shell
python3 prepare_data.py --data_dir "./data/bit vehicle" --output_dir "./dataset_split"
```
## 3. Training
``` shell
python3 model_train.py
```
## 4. Evaluation
# View in time real-time training metrics

``` bash
uvicorn realtime_api:app --host 0.0.0.0 --port 8000
```

Then open `viewer.html` in your web browser by http://127.0.0.1:8000/viewer.html

### 4.1 Model Evaluation
``` shell
python3 model_evaluate.py --model_path "./final_model.pth"
```
## 5. Serve
``` shell
python3 -m uvicorn webView.model_serving:app --host 0.0.0.0 --port 8001 --reload
```
Then open `http://127.0.0.1:8001` in your web browser to access the model serving interface.