| 1,013 | 16 | 368 |
| 下载次数 | 被引频次 | 阅读次数 |
针对番茄叶子病害图像难以采集的问题，提出一种基于生成对抗网络的番茄叶子病害图像增强方法——Residual Network and Wasserstein Attached Spectral Normalization DCGAN(RWS＿DCGAN)。带有谱归一化的残差网络结构构建了新型的生成器模型和判别器模型，引入带有梯度惩罚项的Wasserstein距离。通过实验表明，改进的生成对抗网络RWS＿DCGAN相较于常规增强方法和DCGAN增强方法，能生成病害明显的番茄叶子病害图像，扩充样本数据集，进而能提升分类网络的识别准确率。
Abstract：To solve the problem that tomato leaf disease images are difficult to collect, a Residual network and Wasserstein attached Spectral Normalization DCGAN(RWS＿DCGAN) based on generative adversarial network is proposed. The residual network structure with spectral normalization constructs a new generation model and discriminator model, and introduces the Wasserstein distance with gradient penalty terms. Experimental results show that the improved generative adversarial network RWS＿DCGAN compared with the conventional enhancement method and DCGAN enhancement method, can generate disease-specific tomato leaf disease images, expand the sample data set, and then improve the identification accuracy of the classification network.
[1] 陈雷，袁媛.基于深度迁移学习的农业病害图像识别[J].数据与计算发展前沿，2020,2(2):111-119.
[2] 李辰政，张小俊，朱海涛，等.基于迁移学习的危险行为识别方法研究[J].科学技术与工程，2019,19(16):187-192.
[3] 范世达，马伟荣，姜文博，等.基于深度学习的柑橘黄龙病远程诊断技术初探[J].中国果树，2022(4):76-79.
[4] 孙晓，丁小龙.基于生成对抗网络的人脸表情数据增强方法[J].计算机工程与应用，2020,56(4):115-121.
[5] 曾明昭，高会议，万莉.基于生成对抗网络的葡萄叶片图像数据增强方法[J].仪表技术，2021(5):41-44.
[6] 甘岚，沈鸿飞，王瑶，等.基于改进DCGAN的数据增强方法[J].计算机应用，2021,41(5):1305-1313.
[7] 曹一珉，蔡磊，高敬阳.基于生成对抗网络的基因数据生成方法[J].计算机应用，2022,42(3):783-790.
[8] HEUSEL M,RAMSAUER H,UNTERTHINER T,et al.GANs Trained by a Two Time-scale Update Rule Converge to a Local Nash Equilibrium[J/OL].(2017-06-26)[2022-11-10].https://arxiv.org/abs/1706.08500.
[9] 王万良，李卓蓉.生成式对抗网络研究进展[J].通信学报，2018,39(2):135-148.
[10] GOODFELLOW I J,POUGET-ABADIE J,MIRZA M,et al.Generative Adversarial Nets[C]//27th International Conference on Neural Processing Systems.Montreal:MITPress,2014:2672-2680.
[11] 申瑞彩，翟俊海，侯璎真.多生成器对抗网络[J].河北大学学报(自然科学版),2021,41(6):734-744.
[12] 郭玥秀，杨伟，刘琦，等.残差网络研究综述[J].计算机应用研究，2020,37(5):1292-1297.
[13] 颜贝，张礼，张建林，等.基于残差结构的对抗式网络图像生成方法[J].激光与光电子学进展，2020,57(18):310-317.
[14] GULRAJANI I,AHMED F,ARJOVSKY M,et al.Improved Training of Wasserstein GANs[C]//31st International Conference on Neural Information Processing Systems.New York:ACM,2017:5769-5779.
[15] 李庆旭，王巧华，马美湖.基于生成对抗网络的禽蛋图像数据生成研究[J].农业机械学报.2021,52(2):236-245.
[16] KINGMA D,BA J.Adam:A Method for Stochastic Optimization[J/OL].(2015-06-23)[2022-11-20].https://arxiv.org/abs/1412.6980v6.
[17] SIMONYAN K,ZISSERMAN A.Very Deep Convolutional Networks for Large-scale Image Recognition[J/OL].(2014-09-04)[2022-11-22].https://arxiv.org/abs/1409.1556.
[18] HE K M,ZHANG X Y,REN S Q,et al.Deep Residual Learning for Image Recognition[C]//2016 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE,2016:770-778.
[19] 胡麟苗，张湧.基于生成对抗网络的短波红外-可见光人脸图像翻译[J].光学学报，2020,40(5):75-84.
[20] 陈雪云，黄小巧，谢丽.基于多尺度条件生成对抗网络血细胞图像分类检测方法[J].浙江大学学报(工学版),2021,55(9):1772-1781.
基本信息:
中图分类号:TP391.41;S436.412
引用信息:
[1]祝俊辉,周贤勇,徐明升,等.改进DCGAN数据增强的番茄叶子病害图像识别[J].无线电工程,2023,53(06):1235-1241.
基金信息:
新疆自治区创新人才建设专项自然科学计划(自然科学基金)基金项目(2020D01A132)~~
2022-12-29
2022-12-29
2022-12-29