Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel Features
To solve the missed and wrong detection problems of the object detection model in identifying soybean companion weeds, this paper proposes an enhanced multi-scale channel feature model based on RT-DETR (EMCF-RTDETR). First, we designed a lightweight hybrid-channel feature extraction backbone network...
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-04-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/9/4812 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850279213127434240 |
|---|---|
| author | Hua Yang Yanjie Lyu Yunpeng Jiang Feng Jiang Taiyong Deng Lihao Yu Yuanhao Qiu Hao Xue Junying Guo Zhaoqi Meng |
| author_facet | Hua Yang Yanjie Lyu Yunpeng Jiang Feng Jiang Taiyong Deng Lihao Yu Yuanhao Qiu Hao Xue Junying Guo Zhaoqi Meng |
| author_sort | Hua Yang |
| collection | DOAJ |
| description | To solve the missed and wrong detection problems of the object detection model in identifying soybean companion weeds, this paper proposes an enhanced multi-scale channel feature model based on RT-DETR (EMCF-RTDETR). First, we designed a lightweight hybrid-channel feature extraction backbone network, which consists of a CGF-Block module and a FasterNet-Block module working together, aiming to reduce the amount of computation and the number of parameters while improving the efficiency of feature extraction. Second, we constructed the EA-AIFI module. This module enhances the extraction of detailed features by combining the in-scale feature interaction module with the Efficient Additive attention mechanism. In addition, we designed an Enhanced Multiscale Feature Fusion (EMFF) network structure, which first differentiates the inputs of the three feature layers and then ensures the effective flow between the original and enhanced features of each feature layer by two multiscale feature fusions as well as one diffusion. The experimental results demonstrate that the EMCF-RTDETR model improves the average precision mAP50 and mAP50:95 by 3.3% and 2.2%, respectively, compared to the RT-DETR model, and the FPS is improved by 10%. Moreover, our model outperforms other mainstream detection models in terms of accuracy and speed, revealing its significant potential for soybean weed detection. |
| format | Article |
| id | doaj-art-2d9180b7ac6a4df8a7b0dd6775f90339 |
| institution | OA Journals |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-2d9180b7ac6a4df8a7b0dd6775f903392025-08-20T01:49:09ZengMDPI AGApplied Sciences2076-34172025-04-01159481210.3390/app15094812Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel FeaturesHua Yang0Yanjie Lyu1Yunpeng Jiang2Feng Jiang3Taiyong Deng4Lihao Yu5Yuanhao Qiu6Hao Xue7Junying Guo8Zhaoqi Meng9School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430040, ChinaSchool of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430040, ChinaBiSiCloud (Wuhan) Information Technology Co., Ltd., Wuhan 430024, ChinaSchool of Statistics and Mathematics, Zhongnan University of Economics and Law, Wuhan 430073, ChinaZhengzhou Xinsiqi Technology Co., Ltd., Zhengzhou 450046, ChinaSchool of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430040, ChinaSchool of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430040, ChinaSchool of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430040, ChinaSchool of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430040, ChinaSchool of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430040, ChinaTo solve the missed and wrong detection problems of the object detection model in identifying soybean companion weeds, this paper proposes an enhanced multi-scale channel feature model based on RT-DETR (EMCF-RTDETR). First, we designed a lightweight hybrid-channel feature extraction backbone network, which consists of a CGF-Block module and a FasterNet-Block module working together, aiming to reduce the amount of computation and the number of parameters while improving the efficiency of feature extraction. Second, we constructed the EA-AIFI module. This module enhances the extraction of detailed features by combining the in-scale feature interaction module with the Efficient Additive attention mechanism. In addition, we designed an Enhanced Multiscale Feature Fusion (EMFF) network structure, which first differentiates the inputs of the three feature layers and then ensures the effective flow between the original and enhanced features of each feature layer by two multiscale feature fusions as well as one diffusion. The experimental results demonstrate that the EMCF-RTDETR model improves the average precision mAP50 and mAP50:95 by 3.3% and 2.2%, respectively, compared to the RT-DETR model, and the FPS is improved by 10%. Moreover, our model outperforms other mainstream detection models in terms of accuracy and speed, revealing its significant potential for soybean weed detection.https://www.mdpi.com/2076-3417/15/9/4812RT-DETRweed detectionFasterNetEA attentionfeature fusion |
| spellingShingle | Hua Yang Yanjie Lyu Yunpeng Jiang Feng Jiang Taiyong Deng Lihao Yu Yuanhao Qiu Hao Xue Junying Guo Zhaoqi Meng Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel Features Applied Sciences RT-DETR weed detection FasterNet EA attention feature fusion |
| title | Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel Features |
| title_full | Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel Features |
| title_fullStr | Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel Features |
| title_full_unstemmed | Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel Features |
| title_short | Soybean Weed Detection Based on RT-DETR with Enhanced Multiscale Channel Features |
| title_sort | soybean weed detection based on rt detr with enhanced multiscale channel features |
| topic | RT-DETR weed detection FasterNet EA attention feature fusion |
| url | https://www.mdpi.com/2076-3417/15/9/4812 |
| work_keys_str_mv | AT huayang soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT yanjielyu soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT yunpengjiang soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT fengjiang soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT taiyongdeng soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT lihaoyu soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT yuanhaoqiu soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT haoxue soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT junyingguo soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures AT zhaoqimeng soybeanweeddetectionbasedonrtdetrwithenhancedmultiscalechannelfeatures |