douglasrizzo / dodo_detector_ros

Object detection from images/point cloud using ROS
BSD 3-Clause "New" or "Revised" License
54 stars 11 forks source link

[dodo_detector_ros-1] process has died #1

Closed christenbc closed 5 years ago

christenbc commented 5 years ago

I followed the installation process in this way: Having ROS kinetic installed in Ubuntu v16 with Python, installed:

pip install git+https://github.com/douglasrizzo/dodo_detector.git

roscd; cd ../src; git clone https://github.com/douglasrizzo/dodo_detector_ros.git

Then I use an Intel Sense camera that is integrated in a mobile robot so I change the launcher _detectkinect.launcher accordingly:

<?xml version="1.0"?>
<launch>
    <remap from="/camera/rgb/image_color" to="/camera_floor/driver/color/image_raw"/>
    <remap from="/camera/depth/points" to="/camera_floor/driver/depth/color/points"/>

    <node name="dodo_detector_ros" pkg="dodo_detector_ros" type="detector.py" output="screen">
        <rosparam command="load" file="$(find dodo_detector_ros)/config/main_config.yaml"/>
    </node>
</launch>

Once I make sure that the topics are being published, I execute your package and it gives me the following error:

user@user-VirtualBox:~/catkin_ws/src/dodo_detector_ros/src$ roslaunch dodo_detector_ros detect_kinect.launch 
... logging to /home/christen/.ros/log/c4591196-2ec6-11e9-bd96-94c6911e7b24/roslaunch-user-VirtualBox-3778.log
Checking log directory for disk usage. This may take awhile.
Press Ctrl-C to interrupt
Done checking log file disk usage. Usage is <1GB.

started roslaunch server http://christen-VirtualBox:45683/

SUMMARY
========

PARAMETERS
 * /dodo_detector_ros/detector_type: ssd
 * /dodo_detector_ros/inference_graph: ~/.dodo_detector_...
 * /dodo_detector_ros/label_map: ~/.dodo_detector_...
 * /dodo_detector_ros/sift_database_path: ~/.dodo_detector_...
 * /dodo_detector_ros/sift_min_pts: 10
 * /dodo_detector_ros/ssd_confidence: 0.5
 * /rosdistro: kinetic
 * /rosversion: 1.12.14

NODES
  /
    dodo_detector_ros (dodo_detector_ros/detector.py)

ROS_MASTER_URI=http://192.168.12.20:11311

process[dodo_detector_ros-1]: started with pid [3787]
Traceback (most recent call last):
  File "/home/christen/catkin_ws/src/dodo_detector_ros/src/detector.py", line 12, in <module>
    from dodo_detector.detection import SingleShotDetector, KeypointObjectDetector
  File "/home/christen/.local/lib/python2.7/site-packages/dodo_detector/detection.py", line 11, in <module>
    from object_detection.utils import label_map_util
ImportError: No module named object_detection.utils
[dodo_detector_ros-1] process has died [pid 3787, exit code 1, cmd /home/christen/catkin_ws/src/dodo_detector_ros/src/detector.py /camera_floor/driver/color/image_raw:=/dodo_detector_ros/image_feed /camera_floor/obstacles:=/dodo_detector_ros/pointcloud_feed __name:=dodo_detector_ros __log:=/home/christen/.ros/log/c4591196-2ec6-11e9-bd96-94c6911e7b24/dodo_detector_ros-1.log].
log file: /home/christen/.ros/log/c4591196-2ec6-11e9-bd96-94c6911e7b24/dodo_detector_ros-1*.log
all processes on machine have died, roslaunch will exit
shutting down processing monitor...
... shutting down processing monitor complete
done

What am I doing wrong so that your package does not work?

douglasrizzo commented 5 years ago

My guess is you haven't actually installed the TensorFlow Object Detection API. The steps you mention in your second bullet are incomplete. Please check here. In order to install the API, you need to clone the entire tensorflow/models repo. You also need to at least compile the protobufs and add the libraries to your PYTHONPATH, as mentioned in the installation instructions I linked. That's why you're getting this message:

ImportError: No module named object_detection.utils