There are no reviews yet. Be the first to send feedback to the community and the maintainers!
Repository Details
This project aims at the problem of road surface image denoising and crack recognition by using embedded camera. Using Gaussian filter to blur the image, over the threshold zero processing and the morphology on and off operation are binarization and further denoising. And the crack contour is marked by FAST feature point recognition, which reach for different road crack damage can be identified. After the simulation, the algorithm is transplanted to the Python on Zynq (PYNQ) system to achieve the purpose of crack identification. There are two parts to the innovation: First, using a lower-cost embedded camera to capture photos, and second, using the characteristics of large difference in gray value between crack region and other regions, the crack profile is marked by the way of FAST feature point recognition and use PYNQ to process. It makes the identification system more integrated and portable, and can judge the crack more accurately and reduce the cost.
