Skip to content

Latest commit

 

History

History
191 lines (135 loc) · 8.71 KB

README.md

File metadata and controls

191 lines (135 loc) · 8.71 KB
Raw

<<<<<<< HEAD

HyperRail

temporary message- Please view these branches for latest work on HyperRail from Fall 2021 cs467 course.

Currently branches are working with mock data or demo nodes and will be merged to main when it is ideal.

-web_ui (web app development) https://github.com/Jbruslind/HyperRail/tree/web-ui/WebUI

-ros_sensor_interpretation (self explanatory) https://github.com/Jbruslind/HyperRail/blob/ros_sensor_interpretation/src/hyper_rail/nodes/README.md

-ros_pathplanning (self explanatory) - https://github.com/Jbruslind/HyperRail/tree/ros_pathplanning -merged HyperRail/src from ros_pathplanning to main

-ros_camera_latest (Micasense camera) - https://github.com/Jbruslind/HyperRail/tree/ros_camera_latest

eGreenhouse 2.0

Rewrite for cleaner faster eGreenhouse sensor package

=======

ros_camera_latest Branch

Objective:

  • capture MicaSense images
  • transfer image files from SD of Micasense Camera to external storage drive
  • store file path of images, along with unique id's to SQLlite database for later retrieval
  • delete files from SD card of MicaSense camera once transfered to external storage drive

Camera details:

Micasense RedEdge-MX

HTTP Micasense API

Note: Camera API makes request to 192.168.1.83:80 by default

By default, the ethernet IP address is 192.168.1.83

Please follow guide to connect to camera via ethernet

Setting up enviromental paths

ros_camera_latest

The following should be added to your bash.rc or zsh equivalent

# Goes above all other source calls
export PYTHONPATH="<Absolute Path>/HyperRail/src/hyper_rail/src"

source <Absolute Path>/HyperRail/devel/setup.bash

<<<<<<< HEAD If there is a better way to do this please tell me cause this is really annoying

Basic File Structure Information

  • src
    • ROS programs, contains all nodes, message types, scripts, and source files. Some ROS functionality is dependent on having the database set up and seeded.
  • WebUI
    • Web interface, database, image storage
  • ImageProcessing
    • Code created during image stitching research, includes script for generating test images that can be used with the compositor. Compositor depends on having database setup
  • stable_firmware
    • stable version and firmware of GRBL to control the appropriate driving hardware

 

ROS Programs

Description

The HyperRail is controlled by the following nodes and communicate through the message types listed below.

  • ProgramNode - Recevies program_ids over ProgramFeed topic, gets list of waypoints for the program from the database, coordinates sending movement instructions, data collection instructions, and initiating image stitching.
  • MotionNode - Receives individual waypoints from ProgramNode over MotionService service, generates GCodes to send to the ESP32 controller. Publishes GCodes to GCodeFeed Topic

The following nodes were used during development to test the topics and services. They are available for demonstration, but not used once running with the final hardware.

  • ESPMock - Receives over GCodeFeed, publishes location to MachinePosition
  • CameraMock - Receives requests and sends responses through SensorService
  • SensorMock - Receives requests and sends responses through SensorService

ROS nodes make use of executor classes, db queries, image processing, and constants stored in ~/HyperRail/src/hyper_rail/src

Running the ROS programs

In addition to ROS The following libaraies should be installed, either to a virtual environment or globally

  • pip install opencv-python
  • pip install imutils
  • pip install argparse
  • pip install numpy
  • pip install wheel
  • pip install pgmagick

A Requirements.txt is included in ~/HyperRail/src and can be installed with pip

Related topics and services used by the nodes are located in ../srv and ../msg

Class definitions used in ProgramNode and MotionNode are located in ../src/communication

To run the provided nodes, frist run catkin_make in the root directory.

The included nodes can be run with the given launch file using one of the following commands:

  • roslaunch hyper_rail hyper_rail_test.launch - uses mocks for testing message passing.
  • roslaunch hyper_rail hyper_rail_launch.launch - uses controller nodes to interact with hardware.

Including the argument --screen after the launch file name will print the node output to the terminal.

To run the provided nodes and mocks individually for development, run the following in separate terminals:

  • roscore
  • rosrun hyper_rail ProgramNode
  • rosrun hyper_rail MotionNode
  • rosrun hyper_rail ESPMock
  • rosrun hyper_rail CameraMock
  • rosrun hyper_rail SensorMock
  • roslaunch rosbridge_server rosbridge_websocket.launch

Then send the test program id with rostopic pub /programs hyper_rail/ProgramFeed "program_id: 1"

Image Processing

File paths

Currently it is assumed that images are stored in the public directory of the web-ui ~/HyperRail/WebUI/public/images. Images are stored in using the file structure: /images/{program_run_id}/{image_type}/{image file}

Composite images are stored using the file structure: /images/{program_run_id}/{image_type}_composite{program_run_id}.tif

public/images/
└── 1
    ├── band_1 
    │   ├── 1.tif
    │   ├── 2.tif
    │   ├── 3.tif
    └── band_1_composite1.tif

The image path is set in the settings page of the web-ui and stored in the settings table of the database

Testing with temporary data

  • Note: the database must be migrated and seeded according to the database setup instrutions for database queries to work.

A test image generation script is included as /HyperRail/ImageProcessing/drawing.py This script generates test images which can be used for verifying the compositor crop and layout. To test with this data, generate at least one set of test images in the ~/HyperRail/WebUI/public/images and run the create_test_images() database query, located in ~/HyperRail/src/hyper_rail/src/db_queries.py. This will assign the set of test images to the waypoint runs for program run 1.

Execution

Then the compositor itself can be run on its own using python compositor.py -p {program_run_id} or through the program node. In both cases, ensure that program_run_id is set to 1 if using these demo instructions.

Until the camera is incorporated into the system, the ProgramNode calls compositor.py with a test program_run_id which has image paths associated to it. This is assigned at line 119 in program_executor.py and should be removed when the camera is added.

ToDo:

  • Create robust error handling for lost connection with hardware components
  • Test with hardware
  • Finish sensorAction() function to work with other sensor types. =======

Branch Highlights and Setup

Setup Database and requirements.txt

Install requirements.txt in python enviroment or globally.

In order to test the Micasense Camera, a database is required

Please follow README here.

Note: make sure schema for database is migrated over

Test the Micasense Camera with ROS nodes.

Please follow README located here.

Note: CameraMock node is a ros node that controls the camera and it is called via a ros-service named 'camera_service'.

CameraMock is located here.

Test MicaSense Camera without ROS nodes.

camera_executor.py can be run here.

Note: the explanation for testing the Micasense Camera will be located at bottom of file below classes.

Basic File Structure Information

  • src
    • ROS programs, contains all nodes, message types, scripts, and source files. Some ROS functionality is dependent on having the database set up and seeded.
  • WebUI
    • Web interface, database, image storage
  • ImageProcessing
    • Code created during image stitching research, includes script for generating test images that can be used with the compositor. Compositor depends on having database setup
  • stable_firmware
    • stable version and firmware of GRBL to control the appropriate driving hardware

Future Plans

Needs to be tested with MicaSense camera connected via ethernet.

Currently tested with fake data not utilizing Micasense API.

ros_camera_latest