文件列表:

CMakeLists.txt (7069, 2020-02-08)
Report/ (0, 2020-02-08)
Report/IEEEtran.cls (281957, 2020-02-08)
Report/allweeds.jpg (348668, 2020-02-08)
Report/cnn.jpg (134207, 2020-02-08)
Report/originalplants.jpg (1160397, 2020-02-08)
Report/processedplants.jpg (1015868, 2020-02-08)
Report/weedspray.pdf (2722647, 2020-02-08)
Report/weedspray.tex (10468, 2020-02-08)
images/ (0, 2020-02-08)
images/plants1/ (0, 2020-02-08)
images/plants1/ros_plant0.jpg (191519, 2020-02-08)
images/plants1/ros_plant0_1.jpg (211103, 2020-02-08)
images/plants1/ros_plant0_2.jpg (229798, 2020-02-08)
images/plants2/ (0, 2020-02-08)
images/plants2/train/ (0, 2020-02-08)
images/plants2/train/ros_plant2_0.jpg (142300, 2020-02-08)
images/plants2/train/ros_plant2_1.jpg (287339, 2020-02-08)
images/plants2/train/via_region_data.json (11689, 2020-02-08)
images/plants2/val/ (0, 2020-02-08)
images/plants2/val/ros_plant2_2.jpg (279437, 2020-02-08)
images/plants2/val/via_region_data_val.json (6824, 2020-02-08)
images/plants3/ (0, 2020-02-08)
images/plants3/ros_plant3_0.jpg (332326, 2020-02-08)
images/plants3/ros_plant3_1.jpg (271320, 2020-02-08)
images/preview/ (0, 2020-02-08)
images/preview/allPoints.png (29097, 2020-02-08)
images/preview/weeds_1.png (544407, 2020-02-08)
images/preview/weeds_2.png (656361, 2020-02-08)
images/preview/weeds_3.png (806806, 2020-02-08)
launch/ (0, 2020-02-08)
launch/rosweedspray.launch (554, 2020-02-08)
mask_rcnn/ (0, 2020-02-08)
mask_rcnn/balloon.py (14063, 2020-02-08)
mask_rcnn/cr1.png (892897, 2020-02-08)
mask_rcnn/cr2.png (774040, 2020-02-08)
... ...

# ROS Assignment The objective of the project is to have a simulated robot called Throvald to run across the field of crops and find the bad weeds so that it can spray them. With image processing extracts the middle point of the recognized weed and all of the points are published as a pointCloud. Another node gets these points and desides if it needs to spray and publishes all the collected points so that it will be visualized on Rviz. The final result can be seen . ## Image Processing In order to collect the bad weeds from the images the background had to be removed and then find the weeds. The implementation works with color matching through HSV images, though I also tried with [convolutional neural network (CNN)](https://github.com/bkbilly/ROS-WeedSpray/blob/master/mask_rcnn/README.md). ## Requerements The project requires some dependencies that can be found on the LCAS Repository which can be installed like this: ```bash curl https://raw.githubusercontent.com/LCAS/rosdistro/master/lcas-rosdistro-setup.sh | bash - sudo apt install ros-kinetic-uol-cmp3103m ros-kinetic-uol-cmp9767m-base ros-kinetic-uol-cmp9767m-tutorial ros-kinetic-rqt-image-view ``` ## How to run To run the program there are some requirements. First of all the gazebo simulation must be running with the move_base. It can run everything with these commands: ```bash mkdir -p catkin_ws/src cd catkin_ws git clone https://github.com/bkbilly/weedspray.git src/weedspray catkin_make source devel/setup.bash rosdep install --from-paths . -i -y roslaunch weedspray rosweedspray.launch ``` ## Articles that help * https://www.learnopencv.com/histogram-of-oriented-gradients/ * https://towardsdatascience.com/image-classification-in-10-minutes-with-mnist-dataset-54c35b77a38d * https://www.analyticsvidhya.com/blog/2019/01/build-image-classification-model-10-minutes/

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：