关键词：
病害叶片图像;超分辨率重建;深度学习;LapSRN;DSRNLP;SERS
摘 要：
为了降低农作物病害所带来的损失,借助计算机对农作物病害叶片图像进行图像超分辨率重建具有重要意义。针对基于农作物病害叶片图像的超分辨率重建问题,引入了基于深度学习的农作物病害叶片图像超分辨率重建方法。通过实验将基于深度学习的超分辨率重建方法与两个传统方法 Bicubic和ScSR做了对比,实验结果表明,两个传统方法的PSNR值均未超过15,且SSIM值均未超过0.6。而基于深度学习的网络模型LapSRN、DSRNLP和SERS所得出的PSNR值均接近30,SSIM值均超过了0.6,相比传统方法,性能得到明显提升。
译 名：
Image Super-resolution Reconstruction of Crop Disease Leaves Based on Deep Learning
作 者：
Dai Qiang;Qiao Yan;Cheng Xi;Zhu Cheng;College of Information and Computer,Anhui Agricultural University;College of Computer Science and Engineering,Nanjing University of Science and Technology;
关键词：
diseased leaf images;;super-resolution reconstruction;;deep learning;;LapSRN;;DSRNLP;;SERS
摘 要：
In order to reduce the losses caused by crop diseases,it is of great significance to carry out image super-resolution reconstruction of crop disease leaf images by means of computer. Aiming at the problem of super-resolution reconstruction based on crop disease leaf images,a super-resolution reconstruction method based on deep learning for crop disease leaf images was introduced. Deep resolution reconstruction method based on deep learning was compared with two traditional methods Bicubic and ScSR. The experimental results showed that the PSNR values of the two traditional methods were less than 15,and the SSIM values were less than 0.6. The PSNR values obtained by the deep learning-based network models LapSRN,DSRNLP and SERS were all close to 30,and the SSIM values were all over 0.6. Compared with the traditional method,the performance of super-resolution reconstruction method based on deep learning for crop disease leaf images was significantly improved.
