<!-- TITLE: Vision --> <!-- SUBTITLE: System for performing identification using input from onboard cameras --> # What We Do The computer vision subteam is responsible for developing a system to image the search area designated by the AUVSI SUAS competition and use computer vision algorithms to localize and classify objects of interest to be sent to the ground station front-end and the competition Interoperability server for scoring. # Documentation * [2017-2018 Vision System](/subsystems/vision/docs/2017-2018): information about the vision system we developed for the 2018 AUVSI SUAS competition. * [UAS at UCLA Target Generator](/subsystems/vision/docs/target-generator): how to use our custom target generator script. # UAS Tutorials 1. [Running the Pipeline](/subsystems/vision/projects/running-pipeline) Coming Soon! 2. Image Processing 1. Pillow/PIL 1. [Introduction (`Image` Module)](/subsystems/vision/projects/pillow-intro) 2. [Drawing Shapes, Lines, and Text (`ImageDraw`)](/subsystems/vision/projects/pillow-draw) Coming Soon! 2. OpenCV Coming Soon! # UAS Workshops * [2018 Vision Basics](/fields/workshops/vision) # Resources * [OpenCV Official Documentation](https://docs.opencv.org/3.0-beta/doc/py_tutorials/py_tutorials.html) * [Keras CNN](https://blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html) # Platforms and Libraries Used * [OpenCV](https://github.com/opencv/opencv) for image filtering, manipulation, and segmentation * [Keras](https://github.com/keras-team/keras) for classification * [Darkflow](https://github.com/thtrieu/darkflow) for localization * [Tensorflow](https://github.com/tensorflow/tensorflow) for Keras and Darkflow backend * [Flask-SocketIO](https://github.com/miguelgrinberg/Flask-SocketIO) for the main vision server * [socketIO-client](https://github.com/invisibleroads/socketIO-client) for the vision client workers