Frequency-Based Haze and Rain Removal Network (FHRR-Net) with Deep Convolutional Encoder-Decoder

Removing haze or rain is one of the difficult problems in computer vision applications. On real-world road images, and often occur together, but traditional methods cannot solve this imaging problem. To address simultaneously, we present a robust network-based framework consisting three steps: image decomposition using guided filters, frequency-based removal network (FHRR-Net), restoration based on an atmospheric scattering model predicted transmission maps rain-removed images. We demonstrate FHRR-Net’s capabilities with synthesized Experimental results show that our trained has superior performance test images compared state-of-the-art methods. use PSNR (peak signal-to-noise) SSIM (structural similarity index) indicators to evaluate quantitatively, showing have highest values. Furthermore, through experiments method useful