Panels Detection RTDetr Open-Source Object Detection Model - Accurately Identify Ship Control Panel Equipment

Panels Detection Rtdetr

Developed by cems-official

This model is a fine-tuned object detection model based on PekingU/rtdetr_r101vd_coco_o365, specifically designed for detecting various equipment on ship control panels.

Object Detection

Transformers

Open Source License:Apache-2.0 #Ship Instrument Detection #High-Precision Target Recognition #Multi-Category Panel Localization

Downloads 163

Release Time : 11/19/2024

Model Overview

This model is an object detection model based on the RT-DETR architecture, fine-tuned specifically for recognizing various types of equipment on ship control panels, such as radars, ship management systems, and visual observation devices.

Model Features

Multi-Category Detection

Capable of detecting multiple types of equipment on ship control panels, including radars, ship management systems, and visual observation devices.

Detection of Various Sizes

Effectively detects targets of different sizes, including small, medium, and large equipment.

High-Precision Detection

Achieves high mAP values across multiple categories, particularly excelling in the detection of ship management systems and visual observation devices.

Model Capabilities

Ship Control Panel Equipment Detection

Multi-Category Target Recognition

Detection of Various Sizes

Use Cases

Ship Automation

Ship Control Panel Monitoring

Automatically identifies and monitors the status of various equipment on ship control panels

Can achieve an mAP of 0.56

Marine Equipment Management

Used for automated management and maintenance of ship equipment

Detection mAP for ship management systems can reach 0.76

Maritime Safety

Critical Equipment Monitoring

Monitors the status of critical ship equipment such as radars and thruster controls

Radar detection mAP can reach 0.58-0.70

🚀 panels_detection_rtdetr

This model is a fine - tuned version of PekingU/rtdetr_r101vd_coco_o365 on the None dataset. It can achieve specific evaluation metrics, which is valuable for relevant detection tasks.

📚 Documentation

Model Performance

It achieves the following results on the evaluation set:

Loss: 9.5718
Map: 0.5617
Map 50: 0.6631
Map 75: 0.6137
Map Small: -1.0
Map Medium: 0.3451
Map Large: 0.5935
Mar 1: 0.6546
Mar 10: 0.7877
Mar 100: 0.8058
Mar Small: -1.0
Mar Medium: 0.5802
Mar Large: 0.8672
Map Radar (small): 0.3509
Mar 100 Radar (small): 0.8077
Map Ship management system (small): 0.6748
Mar 100 Ship management system (small): 0.8933
Map Radar (large): 0.5846
Mar 100 Radar (large): 0.8624
Map Ship management system (large): 0.7577
Mar 100 Ship management system (large): 0.9341
Map Ship management system (top): 0.789
Mar 100 Ship management system (top): 0.8356
Map Ecdis (large): 0.3281
Mar 100 Ecdis (large): 0.7652
Map Visual observation (small): 0.585
Mar 100 Visual observation (small): 0.902
Map Ecdis (small): 0.7635
Mar 100 Ecdis (small): 0.8967
Map Ship management system (table top): 0.6306
Mar 100 Ship management system (table top): 0.7882
Map Thruster control: 0.4949
Mar 100 Thruster control: 0.7447
Map Visual observation (left): 0.6062
Mar 100 Visual observation (left): 0.8395
Map Visual observation (mid): 0.7946
Mar 100 Visual observation (mid): 0.8901
Map Visual observation (right): 0.7446
Mar 100 Visual observation (right): 0.8966
Map Bow thruster: 0.2392
Mar 100 Bow thruster: 0.5167
Map Me telegraph: 0.0825
Mar 100 Me telegraph: 0.5143

Training Procedure

Training Hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e - 05
train_batch_size: 4
eval_batch_size: 8
seed: 42
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon = 1e - 08 and optimizer_args = No additional optimizer arguments
lr_scheduler_type: cosine
num_epochs: 7

Training Results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Small	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Small	Mar Medium	Mar Large	Map Radar (small)	Mar 100 Radar (small)	Map Ship management system (small)	Mar 100 Ship management system (small)	Map Radar (large)	Mar 100 Radar (large)	Map Ship management system (large)	Mar 100 Ship management system (large)	Map Ship management system (top)	Mar 100 Ship management system (top)	Map Ecdis (large)	Mar 100 Ecdis (large)	Map Visual observation (small)	Mar 100 Visual observation (small)	Map Ecdis (small)	Mar 100 Ecdis (small)	Map Ship management system (table top)	Mar 100 Ship management system (table top)	Map Thruster control	Mar 100 Thruster control	Map Visual observation (left)	Mar 100 Visual observation (left)	Map Visual observation (mid)	Mar 100 Visual observation (mid)	Map Visual observation (right)	Mar 100 Visual observation (right)	Map Bow thruster	Mar 100 Bow thruster	Map Me telegraph	Mar 100 Me telegraph
14.2599	1.0	699	9.6242	0.4769	0.5404	0.5144	-1.0	0.2274	0.5416	0.5866	0.755	0.7709	-1.0	0.4884	0.8359	0.7408	0.92	0.672	0.8827	0.7054	0.9504	0.8329	0.926	0.7965	0.8692	0.3419	0.9571	0.2734	0.8627	0.1207	0.6933	0.4841	0.7059	0.3541	0.6947	0.5303	0.8961	0.8393	0.9342	0.2629	0.8466	0.1988	0.3583	0.0011	0.0667
8.9356	2.0	1398	9.1941	0.5527	0.6652	0.6044	-1.0	0.3212	0.574	0.6512	0.7882	0.8015	-1.0	0.6608	0.8085	0.6989	0.8862	0.5273	0.8053	0.7683	0.9145	0.7209	0.9073	0.7995	0.8644	0.4929	0.833	0.4034	0.8392	0.5519	0.8333	0.6453	0.8618	0.4221	0.6447	0.5734	0.8474	0.8714	0.8973	0.412	0.8448	0.3154	0.5333	0.0874	0.5095
8.1388	3.0	2097	9.7524	0.535	0.6013	0.5854	-1.0	0.2545	0.574	0.6219	0.7425	0.7612	-1.0	0.538	0.8183	0.6358	0.8292	0.5844	0.8013	0.6721	0.8368	0.7422	0.8829	0.7144	0.8096	0.4904	0.8562	0.7623	0.9078	0.5667	0.89	0.6409	0.7824	0.1853	0.5763	0.5453	0.7789	0.8362	0.9	0.5862	0.9207	0.0384	0.3833	0.0248	0.2619
7.5951	4.0	2796	9.3983	0.5991	0.7001	0.6587	-1.0	0.3745	0.6167	0.6957	0.8036	0.8188	-1.0	0.6611	0.8746	0.603	0.8538	0.626	0.88	0.6211	0.8496	0.8218	0.9382	0.8062	0.8433	0.3917	0.8804	0.6202	0.851	0.8307	0.9433	0.555	0.8147	0.5143	0.8	0.6609	0.8579	0.887	0.9369	0.7174	0.8759	0.2732	0.5333	0.0579	0.4238
7.1786	5.0	3495	9.1194	0.6117	0.7144	0.6689	-1.0	0.3458	0.6476	0.6904	0.8136	0.8324	-1.0	0.6649	0.8777	0.5	0.8538	0.6723	0.8733	0.7272	0.8795	0.778	0.9398	0.7803	0.8385	0.3389	0.8509	0.6484	0.8804	0.7914	0.9433	0.7053	0.8059	0.6257	0.8447	0.5945	0.8658	0.8411	0.9009	0.7812	0.9397	0.2863	0.5792	0.1053	0.4905
7.1386	6.0	4194	9.9394	0.5353	0.634	0.5921	-1.0	0.3062	0.5549	0.6429	0.7691	0.7874	-1.0	0.5638	0.8364	0.3431	0.7631	0.6563	0.8813	0.5789	0.8393	0.6941	0.9236	0.721	0.7712	0.4061	0.8018	0.5685	0.8725	0.7656	0.91	0.5317	0.8	0.5194	0.7684	0.5191	0.8039	0.7994	0.8586	0.6714	0.8793	0.2223	0.4958	0.0333	0.4429
7.0912	7.0	4893	9.5718	0.5617	0.6631	0.6137	-1.0	0.3451	0.5935	0.6546	0.7877	0.8058	-1.0	0.5802	0.8672	0.3509	0.8077	0.6748	0.8933	0.5846	0.8624	0.7577	0.9341	0.789	0.8356	0.3281	0.7652	0.585	0.902	0.7635	0.8967	0.6306	0.7882	0.4949	0.7447	0.6062	0.8395	0.7946	0.8901	0.7446	0.8966	0.2392	0.5167	0.0825	0.5143

Framework Versions

Transformers 4.46.0
Pytorch 2.5.0+cu121
Datasets 3.0.2
Tokenizers 0.20.1

📄 License

This model is released under the apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご