基于混合數據增強的MSWI過(guò)程燃燒狀態(tài)識別

郭海濤; 湯健; 丁海旭; 喬俊飛

doi:10.16383/j.aas.c210843

基于混合數據增強的MSWI過(guò)程燃燒狀態(tài)識別

doi: 10.16383/j.aas.c210843

郭海濤^{1, 2,},
湯健^{1, 2,},
丁海旭^{1, 2,},
喬俊飛^{1, 2,}

1.
北京工業(yè)大學(xué)信息學(xué)部北京 100124
2.
智慧環(huán)保北京實(shí)驗室北京 100124

基金項目: 國家自然科學(xué)基金(62073006, 62021003), 北京市自然科學(xué)基金(4212032, 4192009), 科學(xué)技術(shù)部國家重點(diǎn)研發(fā)計劃(2018YFC1900800-5), 礦冶過(guò)程自動(dòng)控制技術(shù)國家(北京市)重點(diǎn)實(shí)驗室(BGRIMM-KZSKL-2020-02)資助

詳細信息

作者簡(jiǎn)介:
郭海濤：北京工業(yè)大學(xué)信息學(xué)部碩士研究生. 主要研究方向為面向城市固廢焚燒過(guò)程的圖像處理研究. E-mail: guoht@emails.bjut.edu.cn

湯?。罕本┕I(yè)大學(xué)信息學(xué)部教授. 主要研究方向為小樣本數據建模, 城市固廢處理過(guò)程智能控制. 本文通信作者. E-mail: freeflytang@bjut.edu.cn

丁海旭：北京工業(yè)大學(xué)信息學(xué)部博士研究生. 主要研究方向為城市固廢焚燒過(guò)程特征建模與智能控制. E-mail: dinghaixu@emails.bjut.edu.cn

喬俊飛：北京工業(yè)大學(xué)信息學(xué)部教授. 主要研究方向為污水處理過(guò)程智能控制, 神經(jīng)網(wǎng)絡(luò )結構設計與優(yōu)化. E-mail: junfeiq@bjut.edu.cn

計量
- 文章訪(fǎng)問(wèn)數: 922
- HTML全文瀏覽量: 468
- PDF下載量: 162
- 被引次數: 0
出版歷程
- 收稿日期: 2021-09-06
- 錄用日期: 2021-12-02
- 網(wǎng)絡(luò )出版日期: 2022-02-10
- 刊出日期: 2024-03-29

Combustion States Recognition Method of MSWI Process Based on Mixed Data Enhancement

GUO Hai-Tao^{1, 2
,},
TANG Jian^{1, 2
,},
DING Hai-Xu^{1, 2
,},
QIAO Jun-Fei^{1, 2
,}

1.
Faculty of Information Technology, Beijing University of Technology, Beijing 100124
2.
Beijing Laboratory of Smart Environmental Protection, Beijing 100124

Funds: Supported by National Natural Science Foundation of China (62073006, 62021003), Beijing Natural Science Foundation (4212032, 4192009), National Key Research and Development Program of the Ministry of Science and Technology (2018YFC1900800-5), and Beijing Key Laboratory of Process Automation in Mining and Metallurgy (BGRIMM-KZSKL-2020-02)

More Information

Author Bio:
GUO Hai-Tao　Master student at the Faculty of Information Technology, Beijing University of Technology. His main research interest is image processing of municipal solid waste incineration process

TANG Jian　Professor at the Faculty of Information Technology, Beijing University of Technology. His research interest covers small sample data modeling and intelligent control of municipal solid waste treatment process. Corresponding author of this paper

DING Hai-Xu　Ph.D. candidate at the Faculty of Information Technology, Beijing University of Technology. His research interest covers feature modeling and intelligent control of municipal solid waste incineration process

QIAO Jun-Fei　Professor at the Faculty of Information Technology, Beijing University of Technology. His research interest covers intelligent control of wastewater treatment process and structure design and optimization of neural networks

摘要

摘要: 國內城市固廢焚燒(Municipal solid waste incineration, MSWI)過(guò)程通常依靠運行專(zhuān)家觀(guān)察爐內火焰識別燃燒狀態(tài)后再結合自身經(jīng)驗修正控制策略以維持穩定燃燒, 存在智能化水平低、識別結果具有主觀(guān)性與隨意性等問(wèn)題. 由于MSWI過(guò)程的火焰圖像具有強污染、多噪聲等特性, 并且存在異常工況數據較為稀缺等問(wèn)題, 導致傳統目標識別方法難以適用. 對此, 提出一種基于混合數據增強的MSWI過(guò)程燃燒狀態(tài)識別方法. 首先, 結合領(lǐng)域專(zhuān)家經(jīng)驗與焚燒爐排結構對燃燒狀態(tài)進(jìn)行標定; 接著(zhù), 設計由粗調和精調兩級組成的深度卷積生成對抗網(wǎng)絡(luò )(Deep convolutional generative adversarial network, DCGAN)以獲取多工況火焰圖像; 然后, 采用弗雷歇距離(Fréchet inception distance, FID)對生成式樣本進(jìn)行自適應選擇; 最后, 通過(guò)非生成式數據增強對樣本進(jìn)行再次擴充, 獲得混合增強數據構建卷積神經(jīng)網(wǎng)絡(luò )以識別燃燒狀態(tài). 基于某MSWI電廠(chǎng)實(shí)際運行數據實(shí)驗, 表明該方法有效地提高了識別網(wǎng)絡(luò )的泛化性與魯棒性, 具有良好的識別精度.
- 城市固廢焚燒 /
- 深度卷積生成對抗網(wǎng)絡(luò ) /
- 燃燒狀態(tài)識別 /
- 非生成式數據增強 /
- 混合數據增強
Abstract: The municipal solid waste incineration (MSWI) process usually relies on operating experts to observe the flame inside furnace for recognizing the combustion states. Then, by combining the experts＇ own experience to modify the control strategy to maintain the stable combustion. Thus, this manual mode has disadvantages of low intelligence and the subjectivity and randomness recognition results. The traditional methods are difficult to apply to the MSWI process, which has the characteristics of strong pollution, multiple noise, and scarcity of samples under abnormal conditions. To solve the above problems, a combustion states recognition method of MSWI process based on mixed data enhancement is proposed. Firstly, combustion states are labeled by combining the experience of domain experts and the design structure of furnace grate. Next, a deep convolutional generative adversarial network (DCGAN) consisting of two levels of coarse and fine-tuning was designed to acquire multi-situation flame images. Then, the Fréchet inception distance (FID) is used to adaptively select generated samples. Finally, the sample features are enriched at the second time by using non-generative data enhancement strategy, and a convolutional neural network is constructed based on the mixed enhanced data to recognize the combustion state. Experiments based on actual operating data of a MSWI plant show that this method effectively improves the generalization and robustness of the recognition network and has good recognition accuracy.

HTML全文

圖 1 MSWI過(guò)程工藝圖

Fig. 1 Flow chart of MSWI process

下載: 全尺寸圖片幻燈片

圖 2 基于DCGAN數據增強的燃燒狀態(tài)識別策略

Fig. 2 Strategy of combustion state recognition based on DCGAN data enhancement

下載: 全尺寸圖片幻燈片

圖 3 爐排等比例結構示意圖

Fig. 3 Schematic diagram of equal proportion structure of grate

下載: 全尺寸圖片幻燈片

圖 4 燃燒和停爐狀態(tài)圖像標定示意圖

Fig. 4 Image calibration diagram of combustion and shutdown status

下載: 全尺寸圖片幻燈片

圖 5 生成網(wǎng)絡(luò )結構

Fig. 5 Structure of generation network

下載: 全尺寸圖片幻燈片

圖 6 判別網(wǎng)絡(luò )結構

Fig. 6 Structure of discrimination network

下載: 全尺寸圖片幻燈片

圖 7 燃燒線(xiàn)前移

Fig. 7 Combustion line forward

下載: 全尺寸圖片幻燈片

圖 8 燃燒線(xiàn)正常

Fig. 8 Combustion line normal

下載: 全尺寸圖片幻燈片

圖 9 燃燒線(xiàn)后移

Fig. 9 Combustion line back

下載: 全尺寸圖片幻燈片

圖 10 粗調DCGAN迭代過(guò)程中FID對生成燃燒狀態(tài)圖像的評估結果

Fig. 10 Assessment of FID for generating combustion state images during rough DCGAN iteration

下載: 全尺寸圖片幻燈片

圖 11 燃燒線(xiàn)前移的增強圖像

Fig. 11 Expansion results of combustion line forward image

下載: 全尺寸圖片幻燈片

圖 13 燃燒線(xiàn)后移的增強圖像

Fig. 13 Expansion results of combustion line back image

下載: 全尺寸圖片幻燈片

圖 12 燃燒線(xiàn)正常的增強圖像

Fig. 12 Expansion results of combustion line normal image

下載: 全尺寸圖片幻燈片

圖 14 本文所提的非生成式數據增強

Fig. 14 Non-generative data enhancement with the proposed method

下載: 全尺寸圖片幻燈片

圖 15 隨機進(jìn)行的非生成式數據增強

Fig. 15 Non-generative data enhancement with random mode

下載: 全尺寸圖片幻燈片

圖 16 不同生成模型生成的燃燒狀態(tài)圖像

Fig. 16 Combustion state images generated by different generation models

下載: 全尺寸圖片幻燈片

表 1 數據集劃分

Table 1 Dataset partition

數據集	劃分方式	訓練集	驗證集	測試集
A	時(shí)間次序	9 × 8	9 × 1	9 × 1
B	隨機抽樣	9 × 8	9 × 1	9 × 1

下載: 導出CSV

表 2 不同生成模型生成數據的評估結果

Table 2 Evaluation results of data generated by different generation models

方法	評價(jià)指標
方法	FID_min	FID_average	Epoch
GAN	250.00	254.50	10000
LSGAN	58.56	51.94	3000
DCGAN	43.81	49.67	2500
本文方法	36.10	48.51	2500

下載: 導出CSV

表 3 識別模型的性能對比

Table 3 Performance comparison of recognition models

	方法	測試集準確率	測試集損失	驗證集準確率	驗證集損失
方式A	CNN	0.7518±0.00245	0.6046±0.02882	0.6115±0.00212	1.6319±0.11640
	非生成式數據增強+CNN	0.8272±0.00206	0.6504±0.04038	0.7830±0.00183	0.9077±0.03739
	DCGAN數據增強+CNN	0.8000±0.00098	0.8776±0.01063	0.5885±0.00396	1.9024±0.11050
	本文方法	0.8482±0.00105	0.5520±0.01006	0.7269±0.00377	0.9768±0.05797
方式B	CNN	0.8926±0.00105	0.2298±0.00309	0.8519±0.00061	0.2519±0.00167
	非生成式數據增強+CNN	0.9371±0.00184	0.1504±0.00825	0.9704±0.00055	0.1093±0.01037
	DCGAN數據增強+CNN	0.9000±0.00123	0.3159±0.01150	0.8445±0.00207	0.2913±0.00396
	本文方法	0.9407±0.00367	0.2019±0.01498	0.9741±0.00044	0.0699±0.00195

下載: 導出CSV

A1 符號及含義

A1 Symbols and their descriptions

	符號	符號含義
	D	判別器
	G	生成器
	$ V(D,G)$	GAN 原始的目標函數
	${\boldsymbol{z}} $	潛在空間的隨機噪聲
	$ D^*$	固定G 參數, 在$\mathop {\max }\nolimits_D V \left({D,G} \right)$過(guò)程中, D 的最優(yōu)解
	${D_{{\text{JS}}}}$	JS 散度
	${R_{jk}}$	圖像中經(jīng)過(guò)卷積核掃描后的第 j 行第 k 列的結果
	${H_{j - u,k - v}}$	卷積核
	${F_{u,v}}$	圖像
	$X$	燃燒狀態(tài)數據集, 包含前移、正常和后移的數據集, 即燃燒圖像粗調 DCGAN 中判別網(wǎng)絡(luò )輸入值集合$[ { {\boldsymbol{x} }_{{1} } };{ {\boldsymbol{x} }_{{2} } }; $ ${ {\boldsymbol{x} }_{{3} } }; \cdots ;{ {\boldsymbol{x} }_{\rm{a}}} \cdots ]$, 即$ \left[ {{X_{{\rm{real}}}};{X_{{\rm{false}}}}} \right]$
	$ X_{{\rm{FW}}}$	燃燒線(xiàn)前移數據集
	$ X_{{\rm{NM}}}$	燃燒線(xiàn)正常數據集
	$ X_{{\rm{BC}}}$	燃燒線(xiàn)后移數據集
	$ X'_{{\rm{FW}}}$	訓練集燃燒線(xiàn)前移數據集
	$ X'_{{\rm{NM}}}$	訓練集燃燒線(xiàn)正常數據集
	$ X'_{{\rm{BC}}}$	訓練集燃燒線(xiàn)后移數據集
	$ X''_{{\rm{FW}}}$	測試、驗證燃燒線(xiàn)前移數據集
	$ X''_{{\rm{NM}}}$	測試、驗證燃燒線(xiàn)正常數據集
	$ X''_{{\rm{BC}}}$	測試、驗證燃燒線(xiàn)后移數據集
	$ {D_t}(\cdot, \cdot )$	燃燒圖像粗調 DCGAN 子模塊中, 判別網(wǎng)絡(luò )參數為${\theta _{D,t}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ {D_{t+1}}(\cdot, \cdot )$	燃燒圖像粗調 DCGAN 子模塊中, 判別網(wǎng)絡(luò )參數為${\theta _{D,t+1}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ Y_{D,t}$	在燃燒圖像粗調 DCGAN 子模塊中第 t 次博弈訓練判別網(wǎng)絡(luò )的真實(shí)值集合
	$ Y_{G,t}$	在燃燒圖像粗調 DCGAN 子模塊中第 t 次博弈訓練生成網(wǎng)絡(luò )的真實(shí)值集合
	$ loss_{D,t}$	在燃燒圖像粗調 DCGAN 子模塊中第 t 次博弈更新判別網(wǎng)絡(luò )的損失值
	$ loss_{G,t}$	在燃燒圖像粗調 DCGAN 子模塊中第 t 次博弈更新生成網(wǎng)絡(luò )的損失值
	$ X_{{\rm{real}}}$	在燃燒圖像粗調 DCGAN 子模塊中參加博弈的真實(shí)數據
	$ X_{{\rm{false}},t}$	在燃燒圖像粗調 DCGAN 子模塊中參加第 t 次博弈的生成的數據
	$ G_t({\boldsymbol{z}})$	在燃燒圖像粗調 DCGAN 子模塊第 t 次博弈中由隨機噪聲經(jīng)過(guò)生成網(wǎng)絡(luò )得到的虛擬樣本
	${S_{D,t}}$	燃燒圖像粗調 DCGAN 中獲得的判別網(wǎng)絡(luò )的結構參數
	${S_{G,t}}$	燃燒圖像粗調 DCGAN 中獲得的生成網(wǎng)絡(luò )的結構參數
	${\theta _{D,t}}$	在燃燒圖像粗調 DCGAN 子模塊中第 t 次博弈判別網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	${\theta _{G,t}}$	在燃燒圖像粗調 DCGAN 子模塊中第 t 次博弈生成網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	$ X_{{\rm{real}}}^{{\rm{FW}}}$	燃燒線(xiàn)前移精調 DCGAN 子模塊中參加博弈的真實(shí)數據
	$ X_{{\rm{false}},t}^{{\rm{FW}}}$	在燃燒線(xiàn)前移精調 DCGAN 子模塊中參加第 t 次博弈的生成數據
	$ X_{{\rm{real}}}^{{\rm{NM}}}$	燃燒線(xiàn)正常精調 DCGAN 子模塊中參加博弈的真實(shí)數據
	$ X_{{\rm{false}},t}^{{\rm{NM}}}$	在燃燒線(xiàn)正常精調 DCGAN 子模塊中參加第 t 次博弈的生成數據
	$ X_{{\rm{real}}}^{{\rm{BC}}}$	燃燒線(xiàn)后移精調 DCGAN 子模塊中參加博弈的真實(shí)數據
	$ X_{{\rm{false}},t}^{{\rm{BC}}}$	在燃燒線(xiàn)后移精調 DCGAN 子模塊中參加第 t 次博弈的生成數據
	$ D_t^{{\rm{FW}}}(\cdot, \cdot )$	在燃燒線(xiàn)前移精調 DCGAN 子模塊中判別網(wǎng)絡(luò )參數為參數$\theta _{D,t}^{{\text{FW}}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ D_t^{{\rm{NM}}}(\cdot, \cdot )$	在燃燒線(xiàn)正常精調 DCGAN 子模塊中判別網(wǎng)絡(luò )參數為參數$\theta _{D,t}^{{\text{NM}}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ {D}_{t}^{\text{BC}}(\cdot, \cdot ) $	在燃燒線(xiàn)后移精調 DCGAN 子模塊中判別網(wǎng)絡(luò )參數為參數$\theta _{D,t}^{{\text{BC}}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ D_{t+1}^{{\rm{FW}}}(\cdot, \cdot )$	在燃燒線(xiàn)前移精調 DCGAN 子模塊中判別網(wǎng)絡(luò )參數為參數$\theta _{D,t + 1}^{{\text{FW}}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ D_{t+1}^{{\rm{NM}}}(\cdot, \cdot )$	在燃燒線(xiàn)正常精調 DCGAN 子模塊中判別網(wǎng)絡(luò )參數為參數$\theta _{D,t + 1}^{{\text{NM}}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ D_{t+1}^{{\rm{BC}}}(\cdot, \cdot )$	在燃燒線(xiàn)后移精調 DCGAN 子模塊中判別網(wǎng)絡(luò )參數為參數$\theta _{D,t + 1}^{{\text{BC}}}$時(shí), 判別網(wǎng)絡(luò )預測值集合
	$ Y_{D,t}^{{\rm{FW}}}$	燃燒線(xiàn)前移精調 DCGAN 子模塊中第 t 次博弈訓練 D 的真實(shí)值集合
	$ Y_{G,t}^{{\rm{FW}}}$	燃燒線(xiàn)前移精調 DCGAN 子模塊中第 t 次博弈訓練G的真實(shí)值集合
	$ Y_{D,t}^{{\rm{NM}}}$	燃燒線(xiàn)正常精調 DCGAN 子模塊中第 t 次博弈訓練 D 的真實(shí)值集合
	$ Y_{G,t}^{{\rm{NM}}}$	燃燒線(xiàn)正常精調 DCGAN 子模塊中第 t 次博弈訓練G的真實(shí)值集合
	$ Y_{D,t}^{{\rm{BC}}}$	燃燒線(xiàn)后移精調 DCGAN 子模塊中第 t 次博弈訓練 D 的真實(shí)值集合
	$ Y_{G,t}^{{\rm{BC}}}$	燃燒線(xiàn)后移精調 DCGAN 子模塊中第 t 次博弈訓練G的真實(shí)值集合
	$ loss_{D,t}^{{\rm{FW}}}$	燃燒線(xiàn)前移精調 DCGAN 子模塊中第 t 次博弈更新 D 的損失值
	$ loss_{G,t}^{{\rm{FW}}}$	燃燒線(xiàn)前移精調 DCGAN 子模塊中第 t 次博弈更新G的損失值
	$ loss_{D,t}^{{\rm{NM}}}$	燃燒線(xiàn)正常精調 DCGAN 子模塊中第 t 次博弈更新 D 的損失值
	$ loss_{G,t}^{{\rm{NM}}}$	燃燒線(xiàn)正常精調 DCGAN 子模塊中第 t 次博弈更新 G 的損失值
	$ loss_{D,t}^{{\rm{BC}}}$	燃燒線(xiàn)后移精調 DCGAN 子模塊中第 t 次博弈更新 D 的損失值
	$ loss_{G,t}^{{\rm{BC}}}$	燃燒線(xiàn)后移精調 DCGAN 子模塊中第 t 次博弈更新G的損失值
	$\theta _{D,t}^{{\text{FW}}}$	燃燒線(xiàn)前移 DCGAN 子模塊中第 t 次博弈判別網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	$\theta _{G,t}^{{\text{FW}}}$	燃燒線(xiàn)前移 DCGAN 子模塊中第 t 次博弈生成網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	$\theta _{D,t}^{{\text{NM}}}$	燃燒線(xiàn)正常 DCGAN 子模塊中第 t 次博弈判別網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	$\theta _{G,t}^{{\text{NM}}}$	燃燒線(xiàn)正常 DCGAN 子模塊中第 t 次博弈生成網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	$\theta _{D,t}^{{\text{BC}}}$	燃燒線(xiàn)后移 DCGAN 子模塊中第 t 次博弈判別網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	$\theta _{G,t}^{{\text{BC}}}$	燃燒線(xiàn)后移 DCGAN 子模塊中第 t 次博弈生成網(wǎng)絡(luò )更新前的網(wǎng)絡(luò )參數
	${\widehat Y_{{\text{ CNN }},t}}$	燃燒狀態(tài)識別模塊第 t 次更新 CNN 模型預測值集合
	$los{s_{{\text{ CNN }},t}}$	燃燒狀態(tài)識別模塊第 t 次更新 CNN 的損失
	$ \theta _{{\rm{ CNN }},t}$	燃燒狀態(tài)識別模塊第 t 次更新 CNN 的網(wǎng)絡(luò )更新參數
	$ loss$	神經(jīng)網(wǎng)絡(luò )的損失
	${\boldsymbol{x} }_{{a} }$	神經(jīng)網(wǎng)絡(luò )第 a 幅輸入圖像
	$y_a $	第 a 幅輸入圖像輸入神經(jīng)網(wǎng)絡(luò )后的輸出值
	$ D_t(X)$	判別網(wǎng)絡(luò )預測值集合, 即$ {D_t}(\cdot, \cdot )$
	$L $	損失函數
	$\delta_i $	第 i 層的誤差
	$O_i $	第 i 層輸出
	$W_i$	第 i 層的所有權重參數
	$B_i $	第 i 層的所有偏置參數
	$ {\nabla _{{W_{i - 1}}}}$	第$i-1 $層的權重的當前梯度
	$ {\nabla _{{B_{i - 1}}}}$	第$i-1 $層的偏置的當前梯度
	$ {\theta _{D,t}}$	第 t 次判別網(wǎng)絡(luò )的參數
	$ {m _{D,t}}$	第 t 次判別網(wǎng)絡(luò )一階動(dòng)量
	$ {v _{D,t}}$	第 t 次判別網(wǎng)絡(luò )的二階動(dòng)量
	$\alpha $	學(xué)習率
	$\gamma $	很小的正實(shí)數
	$ {\nabla _{D,t}}$	第 t 次判別網(wǎng)絡(luò )參數的梯度
	$\beta_1 $	Adam 超參數
	$\beta_2 $	Adam 超參數
	$ {\eta _{D,t}}$	計算第 t 次的下降梯度
	$ {\widehat m_{D,t}}$	初始階段判別網(wǎng)絡(luò )的第 t 次一階動(dòng)量
	$ {\widehat v_{D,t}}$	初始階段判別網(wǎng)絡(luò )的第 t 次的二階動(dòng)量
	$Y $	神經(jīng)網(wǎng)絡(luò )真值集合
	$ f(X)$	神經(jīng)網(wǎng)絡(luò )預測值集合
	$p $	概率分布
	${p_{\text{r}}}$	真實(shí)圖像的概率分布
	${p_{\text{g}}}$	生成圖像的概率分布
	${p_{\boldsymbol{z}}}$	z 所服從的正態(tài)分布
	Cov	協(xié)方差矩陣

下載: 導出CSV

<button id="qm3rj"><thead id="qm3rj"></thead></button>

亚洲第一网址_国产国产人精品视频69_久久久久精品视频_国产精品第九页

參考文獻(48)

[1]	喬俊飛, 郭子豪, 湯健. 面向城市固廢焚燒過(guò)程的二噁英排放濃度檢測方法綜述. 自動(dòng)化學(xué)報, 2020, 46(6): 1063-1089 Qiao Jun-Fei, Guo Zi-Hao, Tang Jian. Dioxin emission concentration measurement approaches for municipal solid wastes incineration process: A survey. Acta Automatica Sinica, 2020, 46(6): 1063-1089
[2]	Lu J W, Zhang S K, Hai J, Lei M. Status and perspectives of municipal solid waste incineration in China: A comparison with developed regions. Waste Management, 2017, 69: 170-186 doi: 10.1016/j.wasman.2017.04.014
[3]	Kalyani K A, Pandey K K. Waste to energy status in India: A short review. Renewable and Sustainable Energy Reviews, 2014, 31: 113-120 doi: 10.1016/j.rser.2013.11.020
[4]	湯健, 喬俊飛. 基于選擇性集成核學(xué)習算法的固廢焚燒過(guò)程二噁英排放濃度軟測量. 化工學(xué)報, 2018, 70(2): 696-706 Tang Jian, Qiao Jun-Fei. Dioxin emission concentration soft measuring approach of municipal solid waste incineration based on selective ensemble kernel learning algorithm. CIESC Journal, 2019, 70(2): 696-706
[5]	湯健, 王丹丹, 郭子豪, 喬俊飛. 基于虛擬樣本優(yōu)化選擇的城市固廢焚燒過(guò)程二噁英排放濃度預測. 北京工業(yè)大學(xué)學(xué)報, 2021, 47(5): 431-443 Tang Jian, Wang Dan-Dan, Guo Zi-Hao, Qiao Jun-Fei. Prediction of dioxin emission concentration in the municipal solid waste incineration process based on optimal selection of virtual samples. Journal of Beijing University of Technology, 2021, 47(5): 431-443
[6]	湯健, 喬俊飛, 徐喆, 郭子豪. 基于特征約簡(jiǎn)與選擇性集成算法的城市固廢焚燒過(guò)程二噁英排放濃度軟測量. 控制理論與應用, 2021, 38(1): 110-120 Tang Jian, Qiao Jun-Fei, Xu Zhe, Guo Zi-Hao. Soft measuring approach of dioxin emission concentration in municipal solid waste incineration process based on feature reduction and selective ensemble algorithm. Control Theory & Applications, 2021, 38(1): 110-120
[7]	Kolekar K A, Hazra T, Chakrabarty S N. A review on prediction of municipal solid waste generation models. Procedia Environmental Sciences, 2016, 35: 238-244 doi: 10.1016/j.proenv.2016.07.087
[8]	Li X M, Zhang C M, Li Y Z, Zhi Q. The status of municipal solid waste incineration (MSWI) in China and its clean development. Energy Procedia, 2016, 104: 498-503 doi: 10.1016/j.egypro.2016.12.084
[9]	喬俊飛, 段滈杉, 湯健, 蒙西. 基于火焰圖像顏色特征的MSWI燃燒工況識別. 控制工程, 2022, 29(7): 1153-1161 Qiao Jun-Fei, Duan Hao-Shan, Tang Jian, Meng Xi. Recognition of MSWI combustion conditions based on color features of flame images. Control Engineering of China, 2022, 29(7): 1153-1161
[10]	高濟, 何志均. 基于規則的聯(lián)想網(wǎng)絡(luò ). 自動(dòng)化學(xué)報, 1989, 15(4): 318-323 Gao Ji, He Zhi-Jun. The associative network based on rules. Acta Automatica Sinica, 1989, 15(4): 318-323
[11]	Roy S K, Krishna G, Dubey S R, Chaudhuri B B. HybridSN: Exploring 3-D-2-D CNN feature hierarchy for hyperspectral image classification. IEEE Geoscience and Remote Sensing Letters, 2020, 17(2): 277-281 doi: 10.1109/LGRS.2019.2918719
[12]	Ahammad S H, Rajesh V, Rahman Z U, Lay-Ekuakille A. A hybrid CNN-based segmentation and boosting classifier for real time sensor spinal cord injury data. IEEE Sensors Journal, 2020, 20(17): 10092-10101 doi: 10.1109/JSEN.2020.2992879
[13]	Sun Y, Xue B, Zhang M J, Yen G G, Lv J C. Automatically designing CNN architectures using the genetic algorithm for image classification. IEEE Transactions on Cybernetics, 2020, 50(9): 3840-3854 doi: 10.1109/TCYB.2020.2983860
[14]	Zhou P, Gao B H, Wang S, Chai T Y. Identification of abnormal conditions for fused magnesium melting process based on deep learning and multisource information fusion, IEEE Transactions on Industrial Electronics, 2022, 69(3): 3017-3026 doi: 10.1109/TIE.2021.3070512
[15]	張震, 汪斌強, 李向濤, 黃萬(wàn)偉. 基于近鄰傳播學(xué)習的半監督流量分類(lèi)方法. 自動(dòng)化學(xué)報, 2013, 39(7): 1100-1109 Zhang Zhen, Wang Bin-Qiang, Li Xiang-Tao, Huang Wan-Wei. Semi-supervised traffic identification based on affinity propagation. Acta Automatica Sinica, 2013, 39(7): 1100-1109
[16]	王松濤, 周真, 靳薇, 曲寒冰. 基于貝葉斯框架融合的RGB-D圖像顯著(zhù)性檢測. 自動(dòng)化學(xué)報, 2020, 46(4): 695-720 Wang Song-Tao, Zhou Zhen, Jin Wei, Qu Han-Bing. Saliency detection for RGB-D images under Bayesian framework. Acta Automatica Sinica, 2020, 46(4): 695-720
[17]	陶劍文, 王士同. 領(lǐng)域適應核支持向量機. 自動(dòng)化學(xué)報, 2012, 38(5): 797-811 doi: 10.3724/SP.J.1004.2012.00797 Tao Jian-Wen, Wang Shi-Tong. Kernel support vector machine for domain adaptation. Acta Automatica Sinica, 2012, 38(5): 797-811 doi: 10.3724/SP.J.1004.2012.00797
[18]	李強, 王正志. 基于人工神經(jīng)網(wǎng)絡(luò )和經(jīng)驗知識的遙感信息分類(lèi)綜合方法. 自動(dòng)化學(xué)報, 2000, 26(2): 233-239 Li Qiang, Wang Zheng-Zhi. Remote sensing information classification based on artificial neural network and knowledge. Acta Automatica Sinica, 2000, 26(2): 233-239
[19]	羅珍珍, 陳靚影, 劉樂(lè )元, 張坤. 基于條件隨機森林的非約束環(huán)境自然笑臉檢測. 自動(dòng)化學(xué)報, 2018, 44(4): 696-706 Luo Zhen-Zhen, Chen Jing-Ying, Liu Le-Yuan, Zhang Kun. Conditional random forests for spontaneous smile detection in unconstrained environment. Acta Automatica Sinica, 2018, 44(4): 696-706
[20]	郝紅衛, 王志彬, 殷緒成, 陳志強. 分類(lèi)器的動(dòng)態(tài)選擇與循環(huán)集成方法. 自動(dòng)化學(xué)報, 2011, 37(11): 1290-1295 Hao Hong-Wei, Wang Zhi-Bin, Yin Xu-Cheng, Chen Zhi-Qiang. Dynamic selection and circulating combination for multiple classifier systems. Acta Automatica Sinica, 2011, 37(11): 1290-1295
[21]	常亮, 鄧小明, 周明全, 武仲科, 袁野, 楊碩, 等. 圖像理解中的卷積神經(jīng)網(wǎng)絡(luò ). 自動(dòng)化學(xué)報, 2016, 42(9): 1300-1312 Chang Liang, Deng Xiao-Ming, Zhou Ming-Quan, Wu Zhong-Ke, Yuan Ye, Yang Shuo, et al. Convolutional neural networks in image understanding. Acta Automatica Sinica, 2016, 42(9): 1300-1312
[22]	Bai W D, Yan J H, Ma Z Y. Method of flame identification based on support vector machine. Power Engingeering, 2004, 24(4): 548-551
[23]	Sun D, Lu G, Zhou H, Yan Y, Liu S. Quantitative assessment of flame stability through image processing and spectral analysis. IEEE Transactions on Instrumentation and Measurement, 2015, 64(12): 3323-3333 doi: 10.1109/TIM.2015.2444262
[24]	Khan A, Sohail A, Zahoora U, Qureshi A S. A survey of the recent architectures of deep convolutional neural networks. Artificial Intelligence Review, 2020, 53(8): 5455-5516 doi: 10.1007/s10462-020-09825-6
[25]	馮曉碩, 沈樾, 王冬琦. 基于圖像的數據增強方法發(fā)展現狀綜述. 計算機科學(xué)與應用, 2021, 11(2): 370-382 doi: 10.12677/CSA.2021.112037 Feng Xiao-Shuo, Shen Yue, Wang Dong-Qi. A survey on the development of image data augmentation. Computer Science and Application, 2021, 11(2): 370-382 doi: 10.12677/CSA.2021.112037
[26]	Goodfellow I J, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, et al. Generative adversarial nets. In: Proceedings of the 27th International Conference on Neural Information Processing Systems. Montréal, Canada: MIT Press, 2014. 2672?2680
[27]	唐賢倫, 杜一銘, 劉雨微, 李佳歆, 馬藝瑋. 基于條件深度卷積生成對抗網(wǎng)絡(luò )的圖像識別方法. 自動(dòng)化學(xué)報, 2018, 44(5): 855-864 Tang Xian-Lun, Du Yi-Ming, Liu Yu-Wei, Li Jia-Xin, Ma Yi-Wei. Image recognition with conditional deep convolutional generative adversarial networks. Acta Automatica Sinica, 2018, 44(5): 855-864
[28]	劉建偉, 謝浩杰, 羅雄麟. 生成對抗網(wǎng)絡(luò )在各領(lǐng)域應用研究進(jìn)展. 自動(dòng)化學(xué)報, 2020, 46(12): 2500-2536 Liu Jian-Wei, Xie Hao-Jie, Luo Xiong-Lin. Research progress on application of generative adversarial networks in various fields. Acta Automatica Sinica, 2020, 46(12): 2500-2536
[29]	王坤峰, 茍超, 段艷杰, 林懿倫, 鄭心湖, 王飛躍. 生成式對抗網(wǎng)絡(luò )GAN的研究進(jìn)展與展望. 自動(dòng)化學(xué)報, 2017, 43(3): 321-332 Wang Kun-Feng, Gou Chao, Duan Yan-Jie, Lin Yi-Lun, Zheng Xin-Hu, Wang Fei-Yue. Generative adversarial networks: The state of the art and beyond. Acta Automatica Sinica, 2017, 43(3): 321-332
[30]	Yang L, Liu Y H, Peng J Z. An automatic detection and identification method of welded joints based on deep neural network. IEEE Access, 2019, 7: 164952-164961 doi: 10.1109/ACCESS.2019.2953313
[31]	Lian J, Jia W K, Zareapoor M, Zheng Y J, Luo R, Jain D K, et al. Deep-learning-based small surface defect detection via an exaggerated local variation-based generative adversarial network. IEEE Transactions on Industrial Informatics, 2020, 16(2): 1343-1351 doi: 10.1109/TII.2019.2945403
[32]	Niu S L, Li B, Wang X G, Lin H. Defect image sample generation with GAN for improving defect recognition. IEEE Transactions on Automation Science and Engineering, 2020, 17(3): 1611-1622
[33]	Wu X J, Qiu L T, Gu X D, Long Z L. Deep learning-based generic automatic surface defect inspection (ASDI) with pixelwise segmentation. IEEE Transactions on Instrumentation and Measurement, 2021, 70: Article No. 5004010
[34]	Bichler M, Fichtl M, Heidekrüger S, Kohring N, Sutterer P. Learning equilibria in symmetric auction games using artificial neural networks. Nature Machine Intelligence, 2021, 3(8): 687-695 doi: 10.1038/s42256-021-00365-4
[35]	Heusel M, Ramsauer H, Unterthiner T, Nessler B, Hochreiter S. GANs trained by a two time-scale update rule converge to a local Nash equilibrium. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA: Curran Associates Inc., 2017. 6629?6640
[36]	Lucic M, Kurach K, Michalski M, Bousquet O, Gelly S. Are GANs created equal? A large-scale study. In: Proceedings of the 32nd International Conference on Neural Information Processing Systems. Montréal, Canada: Curran Associates Inc., 2018. 698?707
[37]	Radford A, Metz L, Chintala S. Unsupervised representation learning with deep convolutional generative adversarial networks. In: Proceedings of the 4th International Conference on Learning Representations. San Juan, Puerto Rico: 2016.
[38]	Suárez P L, Sappa A D, Vintimilla B X. Infrared image colorization based on a triplet DCGAN architecture. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. Honolulu, USA: IEEE, 2017. 18?23
[39]	Yeh R A, Chen C, Lim T Y, Schwing A G, Hasegawa-Johnson M, Do M N. Semantic image inpainting with deep generative models. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, USA: IEEE, 2017. 5485?5493
[40]	Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks. In: Proceedings of the 26th Annual Conference on Neural Information Processing Systems. Lake Tahoe, USA: Curran Associates Inc., 2012. 1106?1114
[41]	Zeiler M D, Fergus R. Visualizing and understanding convolutional networks. In: Proceedings of the 13th European Conference on Computer Vision. Zurich, Switzerland: Springer, 2014. 818?833
[42]	Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition [Online], available: http://arxiv.org/abs/1409.1556, April 10, 2015
[43]	Szegedy C, Liu W, Jia Y Q, Sermanet P, Reed S, Anguelov D, et al. Going deeper with convolutions. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA: IEEE, 2015. 1?9
[44]	He K M, Zhang X Y, Ren S Q, Sun J. Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, USA: IEEE, 2016. 770?778
[45]	林景棟, 吳欣怡, 柴毅, 尹宏鵬. 卷積神經(jīng)網(wǎng)絡(luò )結構優(yōu)化綜述. 自動(dòng)化學(xué)報, 2020, 46(1): 24-37 Lin Jing-Dong, Wu Xin-Yi, Chai Yi, Yin Hong-Peng. Structure optimization of convolutional neural networks: A survey. Acta Automatica Sinica, 2020, 46(1): 24-37
[46]	Rumelhart D E, Hinton G E, Williams R J. Learning representations by back-propagating errors. Nature, 1986, 323(6088): 533-536 doi: 10.1038/323533a0
[47]	Kingma D P, Ba J. Adam: A method for stochastic optimization. arXiv preprint arXiv: 1412.6980, 2017.
[48]	Mao X D, Li Q, Xie H R, Lau R Y K, Wang Z, Smolley S P. Least squares generative adversarial networks. In: Proceedings of the IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017. 2813?2821