【Visdrone数据集】Visdrone+Cascade_Rcnn_cascade r-cnn模型权重文件下载-优快云博客

本文链接：https://blog.youkuaiyun.com/lxslxslxsllll/article/details/128882807

该文记录了在Visdrone数据集上使用mmdetection框架进行Cascade_Rcnn模型训练的结果。对比了两种不同图像尺寸（416*416和640*640）下，经过10和50个epoch训练后的检测性能，发现分辨率的变化对结果影响不大，各类物体的平均精度和召回率基本保持一致。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

发现img_scale对训练差别不大

Cascade_Rcnn epoch=10 img_scal=416*416

训练代码：mmdetection框架cascade_rcnn_r101_fpn_1x_coco.py

bash ./tools/dist_train.sh  configs/cascade_rcnn/cascade_rcnn_r101_fpn_1x_coco.py  4

分辨率：img_scale=(416, 416)
权重保存文件夹：cascade_rcnn_r101_fpn_1x_coco_img_scale416

测试集

命令：

python tools/test.py configs/cascade_rcnn/cascade_rcnn_r101_fpn_1x_coco.py work_dirs/cascade_rcnn_r101_fpn_1x_coco_img_scale416/epoch_10.pth --options "classwise=True" --eval bbox  --out test.pkl

结果：

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.177
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=1000 ] = 0.311
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=1000 ] = 0.182
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=1000 ] = 0.080
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=1000 ] = 0.282
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=1000 ] = 0.395
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.266
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=300 ] = 0.266
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=1000 ] = 0.266
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=1000 ] = 0.153
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=1000 ] = 0.409
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=1000 ] = 0.493


+------------+-------+-----------------+-------+----------+-------+
| category   | AP    | category        | AP    | category | AP    |
+------------+-------+-----------------+-------+----------+-------+
| pedestrian | 0.089 | people          | 0.034 | bicycle  | 0.048 |
| car        | 0.422 | van             | 0.259 | truck    | 0.246 |
| tricycle   | 0.108 | awning-tricycle | 0.092 | bus      | 0.378 |
| motor      | 0.098 | None            | None  | None     | None  |
+------------+-------+-----------------+-------+----------+-------+
OrderedDict([('bbox_mAP', 0.1774), ('bbox_mAP_50', 0.3111), ('bbox_mAP_75', 0.1825), ('bbox_mAP_s', 0.08), ('bbox_mAP_m', 0.2817), ('bbox_mAP_l', 0.3953), ('bbox_mAP_copypaste', '0.1774 0.3111 0.1825 0.0800 0.2817 0.3953')])

Cascade_Rcnn epoch=50 img_scal=640*640

训练代码：mmdetection框架cascade_rcnn_r101_fpn_1x_coco.py

bash ./tools/dist_train.sh  configs/cascade_rcnn/cascade_rcnn_r101_fpn_1x_coco.py  4

分辨率：img_scale=(640,640)

测试集

命令：

python tools/test.py configs/cascade_rcnn/cascade_rcnn_r101_fpn_1x_coco.py work_dirs/cascade_rcnn_r101_fpn_1x_coco/best_bbox_mAP_epoch_16.pth --options "classwise=True" --eval bbox  --out test.pkl

结果：

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.178                                                                                                         
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=1000 ] = 0.312                                                                                                        
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=1000 ] = 0.183                                                                                                        
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=1000 ] = 0.080                                                                                                        
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=1000 ] = 0.282                                                                                                        
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=1000 ] = 0.398                                                                                                        
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.267                                                                                                         
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=300 ] = 0.267                                                                                                         
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=1000 ] = 0.267                                                                                                        
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=1000 ] = 0.156                                                                                                        
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=1000 ] = 0.411                                                                                                        
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=1000 ] = 0.506                                                                                                        
                                                                                                                                                                                        
                                                                                                                                                                                        
+------------+-------+-----------------+-------+----------+-------+                                                                                                                     
| category   | AP    | category        | AP    | category | AP    |                                                                                                                     
+------------+-------+-----------------+-------+----------+-------+                                                                                                                     
| pedestrian | 0.089 | people          | 0.034 | bicycle  | 0.051 |                                                                                                                     
| car        | 0.421 | van             | 0.258 | truck    | 0.249 |                                                                                                                     
| tricycle   | 0.107 | awning-tricycle | 0.092 | bus      | 0.379 |                                                                                                                     
| motor      | 0.099 | None            | None  | None     | None  |                                                                                                                     
+------------+-------+-----------------+-------+----------+-------+