Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

People-Detection Tutorial - Adit Jha #54

Open
wants to merge 4 commits into
base: master
Choose a base branch
from

Conversation

aditjha
Copy link
Collaborator

@aditjha aditjha commented Sep 22, 2020

tutorial of running people detection on mushr car, summer project by Adit Jha

@aditjha aditjha requested a review from schmittlema September 22, 2020 09:57
Copy link
Member

@schmittlema schmittlema left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good tutorial. There are some changes that need to be made to the code to make it more compatable with our stack. Detection seems OK, The marker didn't move much for me, i'm not conviced that is working properly.

weight: 3
---

<h2> By: <a href=https://www.linkedin.com/in/aditmjha/>Adit Jha (Undergrad)</a></h2>
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you can remove undergrad

### Goal

This tutorial will show you how you can use the RealSense camera on MuSHR cars to detect people and visualize their positions in the car's reference frame.
By the end of this tutorial, you should have a working python package that uses SSD MobileNet V2 to perform
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.


## Getting Started with People Detection

1. Please clone the following repository that contains the python package [`people-detection`](https://github.com/prl-mushr/people_detection).
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

README on repo should follow a format like this: https://github.com/prl-mushr/mushr_rhc
Instead of duplicating the instructions you can just link the tutorial


## Getting Started with People Detection

1. Please clone the following repository that contains the python package [`people-detection`](https://github.com/prl-mushr/people_detection).
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

clone the repo into ../catkin_ws/src/


1. Please clone the following repository that contains the python package [`people-detection`](https://github.com/prl-mushr/people_detection).

This repo contains a `src` directory that holds the code necessary to run object detections and 3D position estimates.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Your repo needs a launch file. It is not a good ros convention to run the python files directly. We need to make this a rospackage too. http://wiki.ros.org/ROS/Tutorials/CreatingPackage


## Getting Started with People Detection

1. Please clone the following repository that contains the python package [`people-detection`](https://github.com/prl-mushr/people_detection).
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

put the clone command in here. cd ~/catkin_ws/src/ && git clone https://github.com/prl-mushr/people_detection.git

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tutorials should be very much like hand holding someone through the process


2. Install the following package used for calculating 3D coordinates:
```
sudo apt-get update
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In front of all commads please put at "$" as is the convention for our tutorials. https://mushr.io/tutorials/quickstart/

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.


- Once rviz is running, add the Robot Model as a topic to see a MuSHR car at (0,0,0), aka the center of the grid.

2. In another terminal window:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you don't have a queue_size param in your publisher which throws a warning. add queue_size=1 in your publisher creation to hide the warning


{{< figure src="/tutorials/people_detection/sample-inference.jpg" width="800" >}}

4. To see Marker Spheres representing the detected people and their relative poses to the car, please go to rviz
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

once I entered the frame an error stopped your code. looks like it is in the ppl_detect.py line 79.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It looks like the projection matrix from image_geometry is empty....

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

it was a subscriber issue, I'll explain after you make the launchfile.

be able to see a green sphere that moves as the person moves in front of the camera.

**NOTE**: Because the inference FPS is slower than the actual RealSense FPS, there is a slight delay in the workflow.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is not taking in the car position, so if the car moves the markers will be off. We should add this. We subscribe to /pf/inferred_pose

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants