Matt 1decd2c3d7 Add solvePNP, 3d tab on the UI, and some other misc bug fixes (#35)
* Rebase solvePNP on master

* added 3D tab minimap and csv reader

* More solvePNP

* Create draw pipe for pnp data

* SolvePNP piping work

* Move sorting into solvepnppipe

* Create calibration pipeline

* Update CalibrateSolvePNPPipeline.java

* add camera tilt angle

* Add calibration slider and snapshot button to 3D view

* Mirror updates in the socket handler

* add 3d calibration mode to the pipeline manager

* created calibration functions in ui and backend

* Start plumbing calibration

* Add snapshot and other handling to the RequestHandler

* added select resolution before starting calibration

* Rename solvePNPPipe to bounding box solve pnp pipe

* Update BoundingBoxSolvePNPPipe.java

* Add Mat serializer and CameraCalibrationConfig

* Begun calibration saving, fixed UI/Backend snapshot count mismatch

* Add (unplumbed) option to set checkerboard size

This will allow users to change the units their calibration is in

* Create chessboard.png

* Fix calibration NPE

* changed string serialization to a json send

* bug fixed cancellation button

* Fix spelling of snapshot in 3d.vue

* Plumb resolution change

* Set resolution during config, start on config serialization

* Update .gitignore

* Config fixes

* Start transition away from cvpipeline3d

* fix NPE on uncalibrated cameras

* clear list on fail

* Fix video mode index error

* ignore getters in camera calibration config

* Create json constructor for jsonmat

* get solvePNP mostly returning sane values

* Fix solvePNP bug and add unit test

* FIx calibration mat truncation

* added capture amount model upload and minimap data

* Standardize on meters in calibration and bounding box

* fix json out of bounds and handle null calibration more gracefully

* don't put text on calibrate image, go back to inches

* convert distance to meters

this means calibration will need to be in inches

* Actually save raw contor

* Update GroupContoursPipe.java

* Add all calibration return to camera capture

* hard code 2019 target

* bugfixed draw2d added fail calib popup, merge end and cancel

added the res index to the calib start

* Clarify error message and draw more fancy rectangles

* Cleanup memory in solvepnp

* re did minimap component

* fix npe if left/right is null

* remove references to 2d

* try-catch running the current pipeline

* Add method to find corners using the harris corner detector

* Possibly fix left/right missmatch

* Fix 3D Tab error

* FIx file permissions, mat serializer adjustments

* fixed mini map for field coordinates

* mini map changes fov

* Update SolvePNPPipe.java

* get rid of target corners

* some memory leak fixes

* fixed mini map location

* added position under minimap

* changed player fov look

* put all targets in the web send

* re did target send to ui added target tables, bugfix calibration

* fixed y position

* Add tilt angle to capture properties

* maybe fix y axis in minimap

* Add square size to onCalibrationEnding

* Possibly add square size to UI

* fix NPE with pitch

* Fix bug with sending multiple targets

* Only instantiate 3d stuff if we are in 3d mode

* Fix array list exceptions

* Fix bug in sort contors

list was truncated too early

* added download chess, tilt setting and ordinal tilt,

* added square size connection

* removed unused code

* Update pom version to 2.1-RELEASE

* Send camera calibrations to UI

* Stream pose list to a LIst

* Only stream necessary parts of the aux list entry

* Make broadcastMessage synchronized to prevent ConcurrentModificationExceptions

* added fps counter changed squaresize steps bug fixes in tables

* bugfix camera settings cam wont change

Authored-by: oriagranat9 <oriagranat9@gmail.com>
2019-12-31 14:53:20 +02:00
2019-12-01 02:07:36 -05:00
2019-11-21 05:32:19 -05:00
2019-12-02 01:46:31 +02:00

Chameleon-Vision

CircleCI CircleCI

Chameleon Vision is free open-source software for FRC teams to use for vision proccesing on their robots.

There instructions are for compiling (contributing) and running the source-code of the project. This is NOT intended for the co-processor setup or your testing PC. To run the program normally (from a build .jar file), take a look at our ReadTheDocs documentation for installation here

These instruction are for the Chameleon Vision's backend/server in Java

To run the UI's sourcecode (optional) see the UI's readme

Hardware

Currently any 64-Bit devices (Windows, Linux and Mac OS) are supported.
32 Bit devices are not supported.

At least one USB camera (supported one is recommended)

Development setup

Prerequisites

  • Java Development Kit 12: Follow the correct instructions for your platform from BellSoft
  • Chameleon-vision source code Clone via a git client or download as zip and extract the source code into a empty folder

For the co-processor(Linux system)

  • Avahi Daemon: sudo apt-get install avahi-daemon avahi-discover avahi-utils libnss-mdns mdns-scan

For the driver station

  • Bonjour Download and install Bonjour from here
  • VC++ Redistributable (Windows only) Download and install this

Importing to IDEA

We recommend the use of Intellij Idea for running the source-code

  1. Import Project

  2. Choose the path to chameleon-server inside the copy of Chameleon-Vision that you cloned or downloaded

  1. Import the project as a Maven project

  1. Under JDK for importer choose the JDK 12 you downloaded earlier
  2. Maven will automatically download the necessary dependencies
  3. Run Main under src/main/java/com/chameleonvision/

Authors

  • Sagi Frimer - initial work - websocket, settings manager, UI

  • Ori Agranat - main coder - project manager, vision loop, UI, websocket, networktables

  • Omer Zipory - developer - vision loop, websocket, networking, documentation, UI

  • Banks Troutman - developer - vision loop, websocket, networking, project structue

  • Matt Morley - developer - vision loop, project structue, documentation, solvePNP

Acknowledgments

License

Usage of Chameleon Vision must fall under all terms of Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International

Description
PhotonVision - FRC Vision Processing
Readme GPL-3.0 639 MiB
Languages
Java 56.5%
C++ 14.9%
Vue 9.2%
Python 7.9%
TypeScript 7.9%
Other 3.4%