在天雨路滑的環境中，雨水影響了影像的品質也導致影像辨識率降低。因此，為了提高雨天的影像辨識率，去雨處理是在影像辨識前，首要處理的事情。在此篇論文中，我們提出了應用卷積神經網路於單圖像去雨演算法。由於雨在影像中所造成之視覺效果，主要是干擾到亮度部分，在本演算法中，為了加快訓練速度且又不影響去雨的效果，因此我們將訓練用之彩色圖，利用HSV彩色模型，將原彩色影像先轉為灰階圖像，以明亮度分量來進行卷積神經網路的訓練。於影像去雨過程中，是將該影像之H與S兩分量先保留，將明亮度的資訊，送入訓練後的卷積神經網路系統中，即可得到去雨後的灰階影像，最後把此灰階影像結合該影像之原有H與S兩分量，轉回成去雨後之彩色圖像。經過實驗後，得到的執行時間上確實比其他方法快，而訓練的時間也有明顯的減少。應用本論文方法進行去雨，以SSIM相似結構性為標準進行評斷，本篇論文的方法與其他方法相比後，結果比其他的高。另外，當雨量稍大時，在影像的視覺上，會感到好像整個影像蒙上一層霧氣，因此，在除完雨後，吾人再使用除霧演算法去進行除霧的動作。其結果之影像與原本無雨影像比較，色澤較為接近，不會產生太大的顏色反差。

In the rainy environment, rain affects the quality of the image and also reduces the image recognition rate. One way to improve the image recognition rate for the rainy images is to remove the rain effects in the image before the processing of image recognition. In the study, a single image de-rain algorithm using a convolutional neural networks are proposed. Due to the effects of rain will mainly contribute the brightness component of the captured image. In order to reduce the training time without affecting the effect of rain, the gray scale image of the input image is applied. For those with rain colored training images, HSV color model is firstly applied to obtain their corresponding grayscale image. Only the brightness component, V are used to train the convolutional neural networks. For each testing image, its H and S components are kept, only the brightness component is applied into the trained convolutional neural networks for derain operation. This result output then combined with its original H and S components to obtain the derained color image. After the experiment, the execution time is indeed faster than other methods, and the training time is also significantly reduced. In this study, the structural similarity index similar structure(SSIM) is used as the evaluation method for comparison. Compared with other methods, the results of the paper are higher than others. Finally, due to the effects of under more rain situation, the image looks like as covered by fog or haze, the dehaze algorithm is applied for these derained image to remove the haze or fog of the image. The colors of result image seem closer to the original one without rain.

[1] Kshitiz Garg and Shree K. Nayar, “Detection and removal of rain from videos.” IEEE Computer Society Conference Computer Vision Pattern Recognition (CVPR’04), Washington, DC, USA, 27 June-2 July, 2004, pp. I-528–I-535.
[2] Xueyang Fu, Jiabin Huang, Xinghao Ding, Yinghao Liao, and John Paisley, “Clearing the skies: A deep network architecture for single-image rain removal.” IEEE Transactions on Image Processing, Vol. 26, No. 6, pp.2944–2956, 2017.
[3] A David Eigen, Dilip Krishnan, and Rob Fergus, “Restoring an image taken through a window covered with dirt or rain.” IEEE International Conference Computer Vision (ICCV), Sydney, NSW, Australia, 1-8 December, 2013, pp. 633–640.
[4] Yu Li, Robby T. Tan, Xiaojie Guo, Jiangbo Lu, and Michael S. Brown, “Rain streak removal using layer priors.” IEEE Conference on Computer Vision Pattern Recognition (CVPR), Las Vegas, NV, USA, 27-30 June, 2016, pp. 2736–2744.
[5] Wenhan Yang, Robby T. Tan, Jiashi Feng, Jiaying Liu, Zongming Guo, and Shuicheng Yan, “Deep joint rain detection and removal from a single image.” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, 21-26 July, 2017, pp.1685–1694.
[6] He Zhang, Vishwanath Sindagi, and Vishal M. Pate, “Image de-raining using a conditional generative adversarial network.” arXiv preprint arXiv:1701.05957, February, 2017.
[7] Gonzalez Woods著, 繆紹綱譯, “數位影像處理”,普林斯頓國際有限公司, pp. 304-323, 2007.
[8] Andrew L. Maas, Awni Y. Hannun and Andrew Y.Ng, “Rectifier Nonlinearities Improve Neural Network Acoustic Models”, in Proceeding of the 30th International Conference on Machine Learning, Vol. 28, Atlanta, Georgia, USA, 17-19 June, 2013.
[9] Xueyang Fu, Jiabin Huang, Delu Zeng, Yue Huang, Xinghao Ding, and John Paisley, “Removing rain from single images via a deep detail network.” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, 21-26 July, 2017, pp.1715–1723.
[10] Photoshop Synthetic rain Image. [online]. Avaiable:
https://www.photoshopessentials.com/photo-effects/photoshop-weather-effects-rain/
[11] Alvy Ray Smith, “Color Gamut Transform Pairs”, Siggraph’78 Proceedings of the 5th annual conference on Computer graphics and interactive techniques, ACM New York, USA, 23-25 August, 1978, pp. 12-19.
[12] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition”, Proceedings of the IEEE, Vol. 86, No. 11, pp. 2278–2324, November 1998.
[13] J. Bossu, N. Hautiere, and J. P. Tarel, “Rain or snow detection in image sequences through use of a histogram of orientation of streaks,” Int. J. Computer Vision, Vol. 93, pp. 348–367, July, 2011.
[14] Vinod Nair and Geoffrey E. Hinton, “Rectified linear units improve restricted boltzmann machines” , in Proceedings of the 27th International Conference on Machine Learning (ICML), Haifa, Israel, 21-24 June, 2010, pp. 807-814.
[15] Qingsong Zhu, Jiaming Mai, and Ling Shao, “A fast single image haze removal algorithm using color attenuation prior”, IEEE Transactions on Image Processing, Vol. 24, No. 11, pp. 3522-3533, Nov., 2015.
[16] Kaiming He, Jian Sun, and Xiaoou Tang, “Single image haze removal using dark channel prior”, IEEE Transaction Pattern Analysis and Machine Intelligence, Vol. 33, No. 12, pp. 2341-2353, Dec., 2011.
[17] Bolun Cai, Xiangmin Xu; Kui Jia, Chunmei Qing, and Dacheng Tao, “DehazeNet: An End-to-End System for Single Image Haze Removal”, IEEE Transactions on Image Processing, Vol. 25, No. 11, pp. 5187-5198, Nov., 2016.
[18] Kaiming He, Jian Sun, and Xiaoou Tang, “Guided image filtering,” IEEE Transactions Pattern Analysis and Machine Intelligence, Vol. 35, No. 6, pp. 1397–1409, Jun., 2013.
[19] Zhou Wang, Alan Conrad Bovik, Hamid Rahim Sheikh, Eero P. Simo-ncelli,“Image Quality Assessment: From Error Visibility to Structural Similarity”, IEEE Transaction on Image Processing, Vol. 13, No. 4, pp. 600-612, Apr., 2004.
[20] Rainy image dataset. [online]. Avaiable:
https://github.com/jinnovation/rainy-image-dataset
[21] Chao Dong, Chen Change Loy, Kaiming He, and Xiaoou Tang, “Image Super-Resolution Using Deep Convolution Network”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 38, No. 2, pp. 295-307, Feb., 2016.
[22] 色階處理, [online]. Avaiable: http://www.flycan.com/article/photoshop/photoshop-levels-110.html