# SmartCar（SSD-mobilenetv2）
Contributions:
1. 基于SSD大框架加入mobilenetv2以实现在嵌入式GPU开发板上的视频流识别
2. 增加 focal loss.
3. 增加detection.py来单纯验证SSD-mobilenetv2算法
4. 提供rs.py来验证D415传感器性能
5. 在detection_D415.py中，加入D415双目摄像头来实现深度预估，以实现物体的动态抓取
6. 提供 visdom可视化，使用python -m visdom.server即可观察收敛
result(train on voc 2007trainval + 2012, test on voc 2007test):
1. ssd-mobielnetv2 (this repo): 70.27%. (without focal loss).
2. ssd-mobielentv1: 68.% (without COCO pretaining), 72.7% (with COCO pretraining) https://github.com/chuanqi305/MobileNet-SSD.
3. ssd-vgg16 (paper): 77.20%.
pretrained model and trained model:
1. 百度网盘: https://pan.baidu.com/s/1RmOPF4jQYpYlE_8E4DifeQ 提取码: f53n
2. Google drive: https://drive.google.com/drive/folders/1JoDYukyWZZ-iWVWPhUDD998cSPB3LeUw?usp=sharing
垃圾分类_基于Jetson+SSD+Mobilenetv2实现的智能小车垃圾分类算法系统_附项目源码_优质项目实战.zip （35个子文件） 
垃圾分类_基于Jetson+SSD+Mobilenetv2实现的智能小车垃圾分类算法系统_附项目源码_优质项目实战 
rs.py 1KB layers __init__.py 48B modules __init__.py 105B multibox_loss.py 6KB l2norm.py 759B __pycache__ l2norm.cpython-36.pyc 1KB multibox_loss.cpython-36.pyc 4KB __init__.cpython-36.pyc 279B box_utils.py 9KB functions __init__.py 97B prior_box.py 2KB detections.py 3KB detection.py 3KB __pycache__ prior_box.cpython-36.pyc 2KB detection.cpython-36.pyc 2KB detections.cpython-36.pyc 2KB __init__.cpython-36.pyc 276B __pycache__ box_utils.cpython-36.pyc 8KB __init__.cpython-36.pyc 200B ssd_mobilenetv2.py 9KB weights ssd_mobilenetv2 mobilenetv2_final.pth 23.14MB mobilenetv2.pth.tar 26.96MB detection.py 6KB data __init__.py 1KB voc0712.py 554B __pycache__ voc0712.cpython-36.pyc 782B __init__.cpython-36.pyc 2KB config.cpython-36.pyc 1KB config.py 1KB detection_D415.py 13KB mobilenetv2.py 7KB test.py 4KB README.md 962B test_videos 
tests.mp4 36MB focal_loss.py 2KB资源评论
__AtYou__
- 粉丝: 3506
- 资源: 2172
