Fraclet Based Deblurring Method Implemented on Pandaboard (Rev A6)
Sonia Saini, Rajiv Kapoor, Manish Kumar Saini · International journal of tomography and simulation · 2020
This paper has proposed the solution of deblurring problem in power insulator images which is a crucial step to be performed before carrying out the fault detection in insulators. The proposed method is based on the decomposition of blurred images using newly designed fractionally delayed wavelet filters (named as fraclets) followed by thresholding of the image coefficients. The reconstructed image effectively removes the blur irrespective of type of blur, e.g., median blur, average blur, bilateral blur and Gaussian blur. The technique has also been implemented on hardware, i.e., Pandaboard (Rev A6) with images captured in laboratory. Two image metrics, i.e., PSNR and SSIM have been considered for quantitatively evaluating the performance of proposed method.