融合GhostNet的YOLOv5垃圾分類方法
電子技術(shù)應(yīng)用
李耀,,胡軍國,,樂楊
浙江農(nóng)林大學(xué) 數(shù)學(xué)與計算機科學(xué)學(xué)院,浙江 杭州 311300
摘要: 垃圾分類是建設(shè)生態(tài)文明的重要一環(huán),為解決重量級模型難以部署移動端設(shè)備的問題,,提出基于YOLOv5網(wǎng)絡(luò)改進的垃圾圖像分類方法,。采用融合GhostNet的主干網(wǎng)絡(luò),,用線性運算代替?zhèn)鹘y(tǒng)卷積運算,,降低了模型的參數(shù)量,提高了模型推理速度,;通過在網(wǎng)絡(luò)中加入改進版通道注意力模塊,,強化重要的通道特征,獲取更多深層次的特征信息,;采用加權(quán)邊界融合方法,,提升檢測框的定位精度。經(jīng)實驗證明,,該方法在自制數(shù)據(jù)集中較原模型的精度提高了8.5%,,參數(shù)量減少了46.7%,平均推理速度提高了1.22 ms,,實現(xiàn)了精度和推理速度的綜合提升,。
中圖分類號:TP391 文獻標(biāo)志碼:A DOI: 10.16157/j.issn.0258-7998.234428
中文引用格式: 李耀,胡軍國,,樂楊. 融合GhostNet的YOLOv5垃圾分類方法[J]. 電子技術(shù)應(yīng)用,,2024,50(1):14-20.
英文引用格式: Li Yao,,Hu Junguo,,Le Yang. YOLOv5 garbage classification method with GhostNet[J]. Application of Electronic Technique,2024,,50(1):14-20.
中文引用格式: 李耀,胡軍國,,樂楊. 融合GhostNet的YOLOv5垃圾分類方法[J]. 電子技術(shù)應(yīng)用,,2024,50(1):14-20.
英文引用格式: Li Yao,,Hu Junguo,,Le Yang. YOLOv5 garbage classification method with GhostNet[J]. Application of Electronic Technique,2024,,50(1):14-20.
YOLOv5 garbage classification method with GhostNet
Li Yao,,Hu Junguo,Le Yang
School of Mathematics and Computer Science,, Zhejiang A&F University,, Hangzhou 311300, China
Abstract: Garbage classification is an important part of building ecological civilization. To solve the problem that heavyweight models are difficult to deploy to mobile devices, an improved garbage image classification method based on YOLOv5 network is proposed. The backbone network fused with GhostNet is used to replace the traditional convolutional operation with linear operation, which reduces the number of parameters of the model and improves the model inference speed. By adding an improved version of channel attention module to the network, the important channel features are strengthened and more deep-level feature information is obtained. The weighted boundary fusion method is used to improve the localization accuracy of the detection frame. It is experimentally demonstrated that the method improves the accuracy by 8.5%, reduces the parameter quantity by 46.7%, and increases the average inference speed by 1.22 ms in the homemade dataset compared with the original model, achieving a comprehensive improvement of accuracy and inference speed.
Key words : garbage classification,;ECA,;GhostNet;YOLOv5
引言
近年來,,隨著我國城市人口的增加,,城市生活垃圾總量增長迅速。據(jù)相關(guān)統(tǒng)計,,我國城市生活垃圾清運量已經(jīng)從2004年的15 509萬噸增長為2021年的24 869萬噸,。自黨的十八大以來,黨中央高度重視生態(tài)文明建設(shè),,垃圾處理問題已經(jīng)成為城市生活中必須解決的問題,,在城市社區(qū)生活中,居民垃圾分類的意識較弱,,且較多依賴于傳統(tǒng)人工分揀,,傳統(tǒng)人工分揀存在耗時長、效率低,、工作環(huán)境差及精準(zhǔn)度較低等問題,。若有效地利用計算機視覺技術(shù)對垃圾進行目標(biāo)檢測,通過對垃圾目標(biāo)的快速定位和精確分類,,將極大地減少人力資源的消耗,,有效地提高垃圾分類效率,為后續(xù)的垃圾回收工作提供支持,,進一步改善城市人居環(huán)境,。
本文詳細內(nèi)容請下載:
http://forexkbc.com/resource/share/2000005827
作者信息:
李耀,胡軍國,,樂楊
(浙江農(nóng)林大學(xué) 數(shù)學(xué)與計算機科學(xué)學(xué)院,,浙江 杭州 311300)
此內(nèi)容為AET網(wǎng)站原創(chuàng),未經(jīng)授權(quán)禁止轉(zhuǎn)載,。