基于改进RT-DETR的轻量级火焰检测算法

doi:10.3969/j.issn.1006-2475.2025.07.015

摘要/Abstract

摘要： 摘要：为了提高火焰检测的精度，同时让模型更加轻量化，提出一种轻量级RT-DETR火焰检测算法。首先，选择EfficientVit作为特征提取网络，减少模型计算量与复杂度。其次，设计一种高效混合编码器，旨在降低模型参数量和计算量的同时，保持检测精度。该编码器由LPE-AIFI模块和CGAFusion模块组成，其中LPE-AIFI模块专注于处理深层特征，而CGAFusion模块通过多尺度特征融合提高模型的检测能力。最后，引入边界框回归损失函数MDPIoU，进一步提高算法精度。实验结果表明，改进后的模型浮点运算数（FLOPs）比原始模型减少了48.8%，参数量减少了43.4%。在达到轻量化的基础上，mAP@0.5值达到88.6%，mAP@0.5:0.95达到67.4%，相较基准模型分别提升了2.2百分点和2.7百分点。

关键词: 关键词：火焰检测, RT-DETR, EfficientViT, 轻量化, 多尺度特征融合

Abstract:
Abstract: In order to improve the accuracy of flame detection and make the model lighter， a lightweight RT-DETR flame detection algorithm is proposed. First， EfficientVit is selected as the feature extraction network to reduce model computation and complexity. Secondly， an efficient hybrid encoder is designed to reduce the number of model parameters and the amount of calculation while maintaining the detection accuracy. The encoder consists of the LPE-AIFI module， which focuses on processing deep features， and the CGAFusion module， which improves the detection capability of the model through multi-scale feature fusion. Finally， the boundary box regression loss function MDPIoU is introduced to further improve the accuracy of the algorithm. The experimental results show that the floating-point operations （FLOPs） of the improved model are reduced by 48.8% and the number of parameters by 43.4% compared with the original model. On the basis of lightweight， mAP@0.5 reaches 88.6% and mAP@0.5:0.95 reaches 67.4%， which are respectively 2.2 percentage points and 2.7 percentage points higher than the benchmark model.

Key words: Key words: flame detection, RT-DETR, EfficientViT, lightweight, multi-scale feature fusion

中图分类号:

中图分类号：TP391.4

吴栋, 范永胜, 桑彬彬. 基于改进RT-DETR的轻量级火焰检测算法[J]. 计算机与现代化, 2025, 0(07): 106-111.

WU Dong, FAN Yongsheng, SAN Binbin. Lightweight Flame Detection Algorithm Based on Improved RT-DETR[J]. Computer and Modernization, 2025, 0(07): 106-111.

参考文献

［1］ GEETHA S， ABHISHEK C S， AKSHAYANAT C S. Machine vision based fire detection techniques: A survey［J］. Fire Technology， 2021，57（2）:591-623.
［2］ LI P， ZHAO W D. Image fire detection algorithms based on convolutional neural networks［J］. Case Studies in Thermal Engineering， 2020，19:100625.
［3］ CHUA L O. CNN: A vision of complexity［J］. International Journal of Bifurcation and Chaos， 1997，7（10）:2219-2425.
［4］ GIRSHICK R， DONAHUE J， DARRELL T， et al. Rich feature hierarchies for accurate object detection and semantic segmentation［C］// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. IEEE， 2014:580-587.
［5］ LIU W， ANGUELOV D， ERHAN D， et al. SSD: Single shot multibox detector［C］// Proceedings of the 14th European Conference on Computer Vision. Springer， 2016:21-37.
［6］ GIRSHICK R. Fast R-CNN［C］// Proceedings of the 2015 IEEE International Conference on Computer Vision. IEEE， 2015:1440-1448.
［7］ ALE L， ZHANG N， LI L Z. Road damage detection using RetinaNet［C］// 2018 IEEE International Conference on Big Data （Big Data）. IEEE， 2018: 5197-5200.
［8］ REDMON J， FARHADI A. YOLO9000: Better， faster， stronger［C］// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. IEEE， 2017:6517-6525.
［9］奚方园，李燕，梅腱，等. 基于改进的YOLOv3火灾检测算法研究［J］. 黑龙江工业学院学报（综合版）， 2023，23（1）:63-70.
［10］徐岩，李永泉，郭晓燕，等. 基于YOLOv3-tiny的火焰目标检测算法［J］. 山东科技大学学报（自然科学版）， 2022，41（6）:95-103.
［11］王海群，张成君，张怡. 基于YOLOv4的轻量级火焰检测算法［J］. 山东科技大学学报（自然科学版）， 2023，42（1）:91-99.
［12］王洪义，孔梅梅，徐荣青. 基于改进YOLOV5的火焰检测算法［J］. 计算机与现代化， 2023（1）:103-107.
［13］谢康康，朱文忠，谢林森，等. 基于改进YOLOv7的火焰烟雾检测算法［J］. 国外电子测量技术， 2023，42（7）:41-49.
［14］李虹，纪任鑫，陈军鹏，等. 基于改进YOLOv5s的森林烟火检测算法研究［J］. 科技创新与应用， 2024，14（5）:7-11.
［15］ ZHAO Y A， LV W Y， XU S L， et al. DETRs beat YOLOs on real-time object detection［C］// Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE， 2024:16965-16974.
［16］ LIU X Y， PENG H W， ZHENG N X， et al. Efficientvit: Memory efficient vision transformer with cascaded group attention［C］// Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE， 2023:14420-14430.
［17］ CHEN Z X， HE Z W， LU Z M. DEA-Net: Single image dehazing based on detail-enhanced convolution and content-guided attention［J］. IEEE Transactions on Image Processing， 2024，33:1002-1015.
［18］ SILIANG M， YONG X. MPDIoU: A loss for efficient and accurate bounding box regression［J］. arXiv preprint arXiv:2307.07662， 2023.
［19］ AVIAN AG. Spark-Detector Computer Vision Project ［DB/OL］. ［2024-07-10］. https://universe.roboflow.com/avian-ag-sd77w/spark-detector/dataset/2.
［20］ REN S Q， HE K M， GIRSHICK R， et al. Faster R-CNN: Towards real-time object detection with region proposal networks［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2016，39（6）:1137-1149.
［21］ CAI Z W， VASCONCELOS N. Cascade R-CNN: High quality object detection and instance segmentation［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2019，43（5）:1483-1498.
［22］ ADARSH P， RATHI P， KUMAR M. YOLO v3-Tiny: Object detection and recognition using one stage improved model［C］// 2020 6th International Conference on advanced Computing and Communication Systems （ICACCS）. IEEE， 2020:687-694.
［23］ WANG D D， HE D J. Channel pruned YOLO V5s-based deep learning approach for rapid and accurate apple fruitlet detection before fruit thinning［J］. Biosystems Engineering， 2021，210:271-281.
［24］ NORKOBIL SAYDIRASULOVICH S， ABDUSALOMOV A， JAMIL M K， et al. A YOLOv6-based improved fire detection approach for smart city environments［J］. Sensors， 2023，23（6）:3161.
［25］ JIANG K L， XIE T Y， YAN R， et al. An attention mechanism-improved YOLOv7 object detection algorithm for hemp duck count estimation［J］. Agriculture， 2022，12（10）:1659.
［26］ WANG Z Y， YUAN G W， ZHOU H， et al. Foreign-object detection in high-voltage transmission line based on improved YOLOV8m［J］. Applied Sciences， 2023，13（23）:12775.
［27］ WANG A， CHEN H， LIU L， et al. YOLOv10: Real-time end-to-end object detection［J］. Advances in Neural Information Processing Systems， 2024，37:107984-108011.
［28］ FENG C J， ZHONG Y J， GAO Y， et al. Tood: Task-aligned one-stage object detection［C］// 2021 IEEE/CVF International Conference on Computer Vision （ICCV）. IEEE， 2021:3490-3499.
［29］ LIU S L， LI F， ZHANG H， et al. DAB-DETR: Dynamic anchor boxes are better queries for DETR［J］. arXiv preprint arXiv:2201.12329， 2022.
［30］ ZHANG H， LI F， LIU S， et al. DINO: DETR with improved denoising anchor boxes for end-to-end object detection［J］. arXiv preprint arXiv:2203.03605， 2022.
［31］ LV W Y， ZHAO Y A， CHANG Q Y， et al. RT-DETRv2: Improved baseline with bag-of-freebies for real-time detection transformer［J］. arXiv preprint arXiv:2407.17140， 2024.
［32］ SAMARTH JAIN. FD1v8n Computer Vision Project ［DB/OL］. ［2024-07-10］.

[1]	江志伟1, 傅晓锦1, 陈文彬1, 江毅晨2. 基于轻量化YOLOv8的无人机对绝缘子缺陷检测[J]. 计算机与现代化, 2025, 0(07): 9-14.
[2]	涂福泉, 戚晏奇, 刘建, 汪曙峰. 基于改进YOLOv8s的金属齿轮表面瑕疵检测算法 [J]. 计算机与现代化, 2025, 0(01): 100-106.
[3]	张思敏, 刘新妹, 殷俊龄, 李宝玲. 基于YOLOv7改进的PCB缺陷检测方法[J]. 计算机与现代化, 2024, 0(12): 45-52.
[4]	张宇1, 2, 黎靖1, 2, 马铭1, 2, 王众祥1, 2, 孙妍1, 2. YOLOLW:一个新的轻量级目标检测模型[J]. 计算机与现代化, 2024, 0(11): 91-98.
[5]	岳有军1, 2, 张远锟1, 赵辉1, 2, 王红君1, 2. 基于多尺度特征与注意力模块的室内场景识别方法[J]. 计算机与现代化, 2024, 0(08): 37-42.
[6]	周永钦, 王勇, 王瑛. 基于多尺度特征及注意力机制的轻量化PCB缺陷检测方法#br#[J]. 计算机与现代化, 2024, 0(02): 88-92.
[7]	欧嘉城, 曾安, 金亮. 基于CP-YOLOX的冷冻电镜图像蛋白质目标检测算法[J]. 计算机与现代化, 2023, 0(11): 113-119.
[8]	罗明杰, 冯开平. 基于沙漏结构与注意力机制的轻量级人脸表情识别方法[J]. 计算机与现代化, 2023, 0(11): 89-94.
[9]	李延满, 王必恒, 赵羚焱. 基于轻量化YOLOv5的安全帽检测[J]. 计算机与现代化, 2023, 0(10): 59-64.
[10]	山雨, 张好鹏, 池静. 基于改进YOLOv4的轻量化车牌检测算法[J]. 计算机与现代化, 2023, 0(07): 99-104.
[11]	黄子平, 黄继风, 周小平. 基于YOLOv3的改进仪表检测算法[J]. 计算机与现代化, 2022, 0(01): 77-84.
[12]	王坭, 王淑营, 史海欧, 袁泉. 基于三角剖分算法的BIM模型高精度显示方法[J]. 计算机与现代化, 2021, 0(09): 57-62.