Software Solutions
Show us the code
Python, Android, Java, C... Is not a poker hand, this are the languages we know, of course will learn together more if required. Implementing your ideas is our passion.
Android Examples
Building apps for Android is one of the easiest way to practice and discover the potential of your smartphone. Our legacy collection gives you an idea that the potential was always there ready to use.
AMIVSI Project
October 2005 Final Year Project (1st Class Hons) written by Amiggy's.
The origins
Analysis and design of an application that detects obstacles, the speed and traffic signals for vehicles with Optical Flow algorithms programmed in C++.
This project is included within the scope of the artificial vision, and more concretely in the estimation of parameters of speed and distance using a sequence of images. In order to obtain these objectives, we will study the present techniques they are most effective, applying them for the attainment of the objectives.
One is to make an application and a library in C/C++ to handle sequences of images caught from the inside of a vehicle in movement. From the treatment of the sequence of images we have to obtain three main targets:
To calculate the speed of movement with respect to the camera, being based on the study of the optical flow of an image we will try to use it in the field of the auto-motion. This calculation is very important because the resolution of the following objectives are based on it.
To consider the free space front to the camera, or depth map from de scene. With a single image we have not the real size of an object, therefore it is not possible compute the distance to this object. It is for that reason why it will be used a sequence of images and an estimation of speed to obtain our objective.
To identify a series of objects previously recognized in previous images, from a detailed list (generated by the SACAT system) of already identified, and to assign a new label to the new objects.
The application uses Sobel filters to detect borders and Optical Flow to detect depths, speed and find objects in motion. It uses a Matrox Imaging Library for optimizing the Optical Flow algorithm.
In the book written, we detail several test and our conclusions were pretty good in terms of accuracy with the following percentages Speed 77,20 % Deph 63,85 % & Labelling 89,54 %
Though this might be consider legacy code, since 2005, with the improvements in the hardware and the new versions of Matrox library you should obtain even greater results. This are common Technics used in Computer Vision specially in automotive so could be a good starting point for anyone interested in autonomous vehicles software.
Source Code written in C++ Git Repository
Academic work book (pdf in Spanish)
PixelJs Roulette
Working in a responsible solution for a casino with an european roulette (Only one 0) where multiple users can play in real time from any device connecting to a server. The top 10 users will be shown in the hall of fame.