This is a hybrid mobile app to record GNSS, Visual and INertial Sensor data.
The app framework used here is Apache Cordova and the app itself is developed using web technologies (HTML5, CSS, JS). The aim is to provide a way to record sensor measurements for developing sensor fusion algorithms using a mobile phone with a common system clock (i.e. asynchronous sensor fusion).
This work was inspired by the GVINS-System from https://github.com/HKUST-Aerial-Robotics/GVINS. My goal is to test their fusion algorithm with my own dataset (and more importantly to ease the dataset creation process!).
https://play.google.com/store/apps/details?id=com.lwtv.gvins_recorder&gl=DE
A user kindly asked about how the recorded sensor data can be used (Thanks to @belkahorry). It was quite embarrasing, that I did not include an example explaining this. Now you can find a plotting script written in Python in this repository (see in the folder "sample_data").
foo@bar:~$ python .\parse-GVINS-Recorder-data.py
{debug output - example for accelerometer}
{'time_ms': 1627315164706, 'x': 0.0017453292519943296, 'y': -0.07155849933176751, 'z': 0.0017453292519943296}
1627315164706.0 ms: 0.0017453292519943296 m/s^2 -0.07155849933176751 m/s^2 0.0017453292519943296 m/s^2
{...}
The script will write the camera images as .png files on the disk and generate the following plots:
(sorry, my phone does not include a magnetometer)
First sensor implementations: GPS, Accelerometer, Gyroscope and Magnetometer. No camera sensor yet.
Added camera sensor implementation. Added error logging to User Interface.
Fixed Camera capturing and added Sensor data export feature. Realized that a file management system with user feedback is needed in the future in order to make debugging easier.
- Storing recordings locally using IndexedDB (should work on browsers and mobile phones)
- Deletion of local recordings
- List recordings in User Interface
- Download recording from IndexedDB to local file system on the phone (does not work the same way on phones and browsers. Mobile is priority but harder to implement)
- Upload recording to my server (show modal dialog asking for permission to publish dataset using Creative-Commons-By-License)
- Real-time visualization of GPS + IMU fused trajectory in 3D viewer
- Offline computation of 3D point clouds from camera images