I am currently writing my proposal for Project #1: Robotics Academy: New Computer Vision Exercises. I just need some guidance on how I should proceed and just a rough idea about what to include in my proposal. I also need to know whether the purpose of this project is to impart a better understanding of computer vision techniques or to solve robotics challenges using computer vision.
I just need to know which techniques I should proceed with.
Here's some of the information I gathered:
Computer Vision Techniques: Simple, Adaptive, Gaussian and Otsu's thresholding. Canny Edge Detection. Blur, Gaussian blur, Laplacian, Laplacian of Gaussian, Sobel and line detection with image convolutions. Size filtering, connected components and contours. Hough transform and Hough transform for circles. Morphological transformations (Erosion, dilation, skeletonization, flood fill). Image Subtraction. Image Averaging. Image Segmentation. Image Pyramids. Foreground Extraction using Active Contours, Harris Corners and Grab-cut algorithm. Training custom object detectors using Histogram of Oriented Gradients and Support Vector Machines.
Computer Vision Exercises: Anti-Aircraft Gun: Using image subtraction, contours and optical flow. Object/Vehicle Speed Estimation: Using centroid tracking and template matching. Follow Object: Using Real-time Object detection and localization using ORB and maybe SIFT (as the patent of SIFT expires in 2020). Disoriented Package Alignment and Skew Correction: Using affine transformations and template matching.
I have completed the programming and academy challenges and I'll attach a copy of the Google form that I filled for evidence.