Skip to content

ROS package to procedural generate randomized fields with rows of plants for Gazebo

License

Notifications You must be signed in to change notification settings

UBCAgroBot/navigation_gazebo

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Virtual Maize Field

Code style: black contributions welcome License: GPL v3

This is a package to procedurally generate randomized fields with rows of plants for Gazebo.

Screenshot of a generated map with maize plants

ROS Distro Gazebo version Branch Build status
Rolling Gazebo ros2-gz ros-rolling-gz
Rolling Gazebo Classic ros2 ros-rolling
Iron Gazebo ros2-gz ros-iron-gz
Iron Gazebo Classic ros2 ros-iron
Humble Ignition Gazebo ros2-ign ros-humble-ign
Humble Gazebo Classic ros2 ros-humble
Noetic Gazebo Classic main ros-noetic

Installation

Clone this repository and build the workspace. Additional you'll need the following packages:

rosdep install virtual_maize_field

This package has been tested on ROS 2 Humble, Iron and Rolling.

Generating new maize field worlds

This package includes a script (virtual_maize_field/generate_world.py) that can generate randomized agricultural worlds. All parameters are optional and have default values. All comma separated arguments can be scaler as well.

You can call the script using

ros2 run virtual_maize_field generate_world

The resulting file will be placed in $ROS_HOME/virtual_maize_field/generated.world.

You can use this script by one of the defined config files or specifying the parameters below:

Click to show all possible arguments
usage: generate_world [-h] [--row_length ROW_LENGTH]
                    [--rows_curve_budget ROWS_CURVE_BUDGET]
                    [--row_width ROW_WIDTH] [--rows_count ROWS_COUNT]
                    [--row_segments [ROW_SEGMENTS [ROW_SEGMENTS ...]]]
                    [--row_segment_straight_length_min ROW_SEGMENT_STRAIGHT_LENGTH_MIN]
                    [--row_segment_straight_length_max ROW_SEGMENT_STRAIGHT_LENGTH_MAX]
                    [--row_segment_sincurved_offset_min ROW_SEGMENT_SINCURVED_OFFSET_MIN]
                    [--row_segment_sincurved_offset_max ROW_SEGMENT_SINCURVED_OFFSET_MAX]
                    [--row_segment_sincurved_length_min ROW_SEGMENT_SINCURVED_LENGTH_MIN]
                    [--row_segment_sincurved_length_max ROW_SEGMENT_SINCURVED_LENGTH_MAX]
                    [--row_segment_curved_radius_min ROW_SEGMENT_CURVED_RADIUS_MIN]
                    [--row_segment_curved_radius_max ROW_SEGMENT_CURVED_RADIUS_MAX]
                    [--row_segment_curved_arc_measure_min ROW_SEGMENT_CURVED_ARC_MEASURE_MIN]
                    [--row_segment_curved_arc_measure_max ROW_SEGMENT_CURVED_ARC_MEASURE_MAX]
                    [--row_segment_island_radius_min ROW_SEGMENT_ISLAND_RADIUS_MIN]
                    [--row_segment_island_radius_max ROW_SEGMENT_ISLAND_RADIUS_MAX]
                    [--ground_resolution GROUND_RESOLUTION]
                    [--ground_elevation_max GROUND_ELEVATION_MAX]
                    [--ground_headland GROUND_HEADLAND]
                    [--ground_ditch_depth GROUND_DITCH_DEPTH]
                    [--plant_spacing_min PLANT_SPACING_MIN]
                    [--plant_spacing_max PLANT_SPACING_MAX]
                    [--plant_height_min PLANT_HEIGHT_MIN]
                    [--plant_height_max PLANT_HEIGHT_MAX]
                    [--plant_radius PLANT_RADIUS]
                    [--plant_radius_noise PLANT_RADIUS_NOISE]
                    [--plant_placement_error_max PLANT_PLACEMENT_ERROR_MAX]
                    [--plant_mass PLANT_MASS]
                    [--hole_prob [HOLE_PROB [HOLE_PROB ...]]]
                    [--hole_size_max [HOLE_SIZE_MAX [HOLE_SIZE_MAX ...]]]
                    [--crop_types [CROP_TYPES [CROP_TYPES ...]]]
                    [--litters LITTERS]
                    [--litter_types [LITTER_TYPES [LITTER_TYPES ...]]]
                    [--weeds WEEDS]
                    [--weed_types [WEED_TYPES [WEED_TYPES ...]]]
                    [--ghost_objects GHOST_OBJECTS]
                    [--location_markers LOCATION_MARKERS]
                    [--load_from_file LOAD_FROM_FILE] [--seed SEED]
                    [--show_map]
                    [{fre21_task_2_mini,fre21_task_2_fast,fre21_task_1_mini,fre22_task_mapping_mini,fre22_task_navigation,fre21_task_3_fast,fre22_task_mapping,fre21_task_4_fast,fre21_task_2,fre21_task_3_mini,fre22_task_mapping_fast,fre21_task_1,fre21_task_3,fre22_task_navigation_fast,fre21_task_1_fast,fre21_task_4,fre22_task_navigation_mini,fre21_task_4_mini}]

Generate a virtual maize field world for Gazebo.

positional arguments:
{fre21_task_2_mini,fre21_task_2_fast,fre21_task_1_mini,fre22_task_mapping_mini,fre22_task_navigation,fre21_task_3_fast,fre22_task_mapping,fre21_task_4_fast,fre21_task_2,fre21_task_3_mini,fre22_task_mapping_fast,fre21_task_1,fre21_task_3,fre22_task_navigation_fast,fre21_task_1_fast,fre21_task_4,fre22_task_navigation_mini,fre21_task_4_mini}
                      Config file name in the config folder

optional arguments:
-h, --help            show this help message and exit
--row_length ROW_LENGTH
                      default_value: 12.0
--rows_curve_budget ROWS_CURVE_BUDGET
                      default_value: 1.5707963267948966
--row_width ROW_WIDTH
                      default_value: 0.75
--rows_count ROWS_COUNT
                      default_value: 6
--row_segments [ROW_SEGMENTS [ROW_SEGMENTS ...]]
                      default_value: ['straight', 'curved']
--row_segment_straight_length_min ROW_SEGMENT_STRAIGHT_LENGTH_MIN
                      default_value: 0.5
--row_segment_straight_length_max ROW_SEGMENT_STRAIGHT_LENGTH_MAX
                      default_value: 1
--row_segment_sincurved_offset_min ROW_SEGMENT_SINCURVED_OFFSET_MIN
                      default_value: 0.5
--row_segment_sincurved_offset_max ROW_SEGMENT_SINCURVED_OFFSET_MAX
                      default_value: 1.5
--row_segment_sincurved_length_min ROW_SEGMENT_SINCURVED_LENGTH_MIN
                      default_value: 3
--row_segment_sincurved_length_max ROW_SEGMENT_SINCURVED_LENGTH_MAX
                      default_value: 5
--row_segment_curved_radius_min ROW_SEGMENT_CURVED_RADIUS_MIN
                      default_value: 3.0
--row_segment_curved_radius_max ROW_SEGMENT_CURVED_RADIUS_MAX
                      default_value: 10.0
--row_segment_curved_arc_measure_min ROW_SEGMENT_CURVED_ARC_MEASURE_MIN
                      default_value: 0.3
--row_segment_curved_arc_measure_max ROW_SEGMENT_CURVED_ARC_MEASURE_MAX
                      default_value: 1.0
--row_segment_island_radius_min ROW_SEGMENT_ISLAND_RADIUS_MIN
                      default_value: 1.0
--row_segment_island_radius_max ROW_SEGMENT_ISLAND_RADIUS_MAX
                      default_value: 3.0
--ground_resolution GROUND_RESOLUTION
                      default_value: 0.02
--ground_elevation_max GROUND_ELEVATION_MAX
                      default_value: 0.2
--ground_headland GROUND_HEADLAND
                      default_value: 2.0
--ground_ditch_depth GROUND_DITCH_DEPTH
                      default_value: 0.3
--plant_spacing_min PLANT_SPACING_MIN
                      default_value: 0.13
--plant_spacing_max PLANT_SPACING_MAX
                      default_value: 0.19
--plant_height_min PLANT_HEIGHT_MIN
                      default_value: 0.3
--plant_height_max PLANT_HEIGHT_MAX
                      default_value: 0.6
--plant_radius PLANT_RADIUS
                      default_value: 0.3
--plant_radius_noise PLANT_RADIUS_NOISE
                      default_value: 0.05
--plant_placement_error_max PLANT_PLACEMENT_ERROR_MAX
                      default_value: 0.02
--plant_mass PLANT_MASS
                      default_value: 0.3
--hole_prob [HOLE_PROB [HOLE_PROB ...]]
                      default_value: [0.06, 0.06, 0.04, 0.04, 0.0, 0.0]
--hole_size_max [HOLE_SIZE_MAX [HOLE_SIZE_MAX ...]]
                      default_value: [7, 5, 5, 3, 0, 0]
--crop_types [CROP_TYPES [CROP_TYPES ...]]
                      default_value: ['maize_01', 'maize_02']
--litters LITTERS     default_value: 0
--litter_types [LITTER_TYPES [LITTER_TYPES ...]]
                      default_value: ['ale', 'beer', 'coke_can',
                      'retro_pepsi_can']
--weeds WEEDS         default_value: 0
--weed_types [WEED_TYPES [WEED_TYPES ...]]
                      default_value: ['nettle', 'unknown_weed', 'dandelion']
--ghost_objects GHOST_OBJECTS
                      default_value: False
--location_markers LOCATION_MARKERS
                      default_value: False
--load_from_file LOAD_FROM_FILE
                      default_value: None
--seed SEED           default_value: -1
--show_map            Show map after generation.

Sample Worlds

In the config folder, config files to generate sample worlds are located. The parameters are chosen to match the task description

Worlds for the Field Robot Event 2022:

Name Description
fre22_task_navigation Task navigation, curved rows that get more difficult (eg. have more and larger holes) to the left
fre22_task_mapping Task mapping, field with random holes, bottles and weeds spread throughout the field. The cans, bottles and weeds have no collision box and are static.
This world needs dandelion models which were only distributed among competitors of the Field Robot Event 2022. They are not uploaded to Github because they cannot be open-sourced. If you don't have access to these models, check out the fre21_task_3 worlds.

Worlds for the Field Robot Event 2021:

Name Description
fre21_task_1 Task 1, curved rows without holes
fre21_task_2 Task 2, straight rows with holes
fre21_task_3 Task 3, similar crop rows as in task_2 but with cans, bottles and weeds spread throughout the field. The cans, bottles and weeds have no collision box and are static.
fre21_task_4 Task 4, similar crop rows as in task_2 but with cans, bottles and weeds spread throughout the field. The cans, bottles and weeds have a collision box and can be picked up.

You can use these config files when generating worlds, e.g.:

ros2 run virtual_maize_field generate_world fre22_task_navigation_mini

Launching and using generated worlds

The launch file to launch the worlds is called simulation.launch. You can launch the launch file by running ros2 launch virtual_maize_field simulation.launch.py. By default the launch file will launch generated_world.world. You can launch any world by using the world_name arg. e.g. ros2 launch virtual_maize_field simulation.launch.py world_name:=simple_row_level_1.world. The generated world will be saved in $ROS_HOME/virtual_maize_field (usually, this will be ~/.ros/virtual_maize_field).

To add your own robot in the world, use the generated robot_spawner.launch.py. This launches your robot at the correct position in the generated world. Your launch file to launch your robot should look like (replace <<robot_name>> with your robot name):

from __future__ import annotations

from launch import LaunchDescription
from launch.actions import IncludeLaunchDescription
from launch.launch_description_sources import PythonLaunchDescriptionSource

from virtual_maize_field import get_spawner_launch_file


def generate_launch_description() -> LaunchDescription:
    robot_spawner_launch = IncludeLaunchDescription(
        PythonLaunchDescriptionSource([get_spawner_launch_file()]),
        launch_arguments={"robot_name": "<<robot_name>>"}.items(),
    )

    return LaunchDescription([robot_spawner_launch])

Use the function get_driving_pattern_file() to get the path of the generated driving pattern and get_markers_file to get the path of the generated markers:

from __future__ import annotations

from csv import reader
from pathlib import Path

from virtual_maize_field import get_driving_pattern_file, get_markers_file

def read_driving_pattern() -> None:
    driving_pattern = Path(get_driving_pattern_file()).read_text("utf-8")
    print(f"The driving pattern is {driving_pattern}")

def read_markers_file() -> None:
    with open(get_markers_file(), "r") as f:
        reader = csv.reader(f)

        # Skip header
        next(reader)

        for row in reader:
            print(f"Position {row[2]}: x={float(row[0]):.3f} y={float(row[1]):.3f}")

License

Virtual Maize Field is copyright (C) 2021 Farm Technology Group of Wageningen University & Research and Kamaro Engineering e.V. and licensed under License: GPL v3.

Models

Name Path Copyright License
Maize 01 models/maize_01/ 2021 Kamaro Engineering e.V. License: CC BY-SA 4.0
Maize 02 models/maize_02/ 2021 Kamaro Engineering e.V. License: CC BY-SA 4.0
Stone 01 models/stone_01/ 2020 Andrea Spognetta License: CC BY-NC 4.0
Stone 02 models/stone_02/ 2014 Sascha Henrichs License: CC BY 4.0
Ale models/ale/ 2017 elouisetrewartha License: CC BY-NC 4.0
Beer models/beer/ ? Maurice Fallon ?
Coke Can models/coke_can/ ? John Hsu ?
Nettle models/nettle/ 2019 LadyIReyna License: CC BY-NC 4.0
Retro Pepsi Can models/retro_pepsi_can/ 2018 FWTeastwood License: CC BY 4.0
Unknown Weed models/unknown_weed/ 2016 aaron_nerlich License: CC BY 4.0

Textures

Name Path Copyright License
grass Media/models/materials/textures/ 2018 CC0Textures.com License: CC0-1.0

About

ROS package to procedural generate randomized fields with rows of plants for Gazebo

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%