Visualization and Editing of Camera Tiepoints, Orientations, and Residuals
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.
You will need to install this software beforehand:
Node.js v16.15.1
npm v8.11.0
To install Node.js and npm, you can use nvm:
- Install nvm from GitHub. Running this command will attempt to add the nvm path to the user profile. This can fail and a warning will be outputted. To resolve the issue, please look at the GitHub.
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.1/install.sh | bash- Install the correct version of Node.js.
nvm install 16.15.1- Set the default Node.js version.
nvm alias default 16.15.1- Restart your terminal and confirm your Node.js version.
node --version
> v16.15.1- Clone the repository.
git clone [email protected]:vis-program/vector.git- Switch to the project directory and install the necessary dependencies.
cd vector
npm install- Run the application.
npm run devVECTOR loads files using the File System Access API available in modern browsers. VECTOR introduces a Loader interface that can be implemented to parse a File object into XML, PNG, VICAR, etc...
Each implementation of Loader includes an EXTENSIONS property that defines what extensions a Loader supports. For example, this is the XMLLoader:
export default class XMLLoader extends Loader {
static EXTENSIONS = ['xml', 'tie', 'tpt', 'nav'];
...
}This is useful when files have custom extensions that are non-standard. If you need to support a new file extension, you can update the EXTENSIONS property.
Anyone can implement a new Loader by extending the abstract base class Loader and satisfying the type constraints. You will also need to update the Landing route to handle your new loader.
export default class MyLoader extends Loader {
static EXTENSIONS: string[] = ['kazi', 'jawad'];
static async load(...): Promise<...> {
...
}
}If a method is not implemented an Error will be thrown when that method is called inside VECTOR.
However, not every method needs an implementation. If you are not exporting to a file type, the write method does not need to be implemented.
VECTOR requires camera, track, and image information to be uploaded by the user.
To allow for flexibility in the data processing, VECTOR exposes a Format interface that can be implemented to translate various formats into VECTOR's data structures.
VECTOR does not handle different coordinate systems, so tracks and cameras are expected to be in the same coordinate system.
The VISOR format is currently used at NASA's Jet Propulsion Laboratory for the Mars 2020 mission.
It defines a navigation.xml and tiepoints.xml alongside a directory of images that match against a unique_id attribute defined inside the tiepoints.xml.
The VECTOR format is specifically designed for VECTOR. Therefore, it involves less processing time and is structured exactly like VECTOR's internal data structures.
It defines a cameras.xml and tracks.xml alongside a directory of images, where the cameras.xml includes the filename for the image each camera maps to.
Anyone can implement a new Format by extending the abstract base class Format and satisfying the type constraints. You will also need to update the Landing route to handle your new format.
class MyFormat extends Format {
static async processTracks(...): Promise<Track[]> {
...
}
static async processCameras(_: unknown): Promise<Camera[]> {
...
}
...
}If a method is not implemented an Error will be thrown when that method is called inside VECTOR.
However, not every method needs an implementation.
For example, the mapImages method is only needed if image names are not available when cameras are being processed. This is the case for the VISOR format, but you will notice the VECTOR format does not implement this method.
Likewise, if you are not planning on exporting file information, the export methods do not need to be implemented.
VECTOR supports a CameraModel interface that allows anyone to implement their own geoemtric camera model for their specific use cases. The CameraModel interface handles geometric details like frustum visualization or ray-image projection.
The CAHVORE camera model is used at NASA's Jet Propulsion Laboratory and is the primary reference implementation.
Anyone can implement a new Format by extending the abstract base class Format and satisfying the type constraints.
class MyCameraModel extends CameraModel {
static ID = 'KaziJawad';
...
}If a method is not implemented an Error will be thrown when that method is called inside VECTOR.
The ID property can be used inside the Format interface to validate against different camera models.
The VECTOR format specifically has a model attribute that matches against the ID string.
This static property might not always be needed. For example, it is not used in the VISOR format because that always uses the CAHVORE camera model.
An alternative to using the Loader and Format interfaces for data processing is converting to VECTOR's format and use that as input into VECTOR.
This is useful for people inexperienced with JavaScript/TypeScript and are more comfortable writing a script.
We have provided a reference implemention in Python that converts the VISOR format to the VECTOR format without external dependencies inside the cmd folder.
For further reference, a presentation on VECTOR can be found here.
Kazi Jawad, Racquel Fygenson, Isabel Li, Mauricio Hess-Flores, François Ayoub, Robert Deen, Scott Davidoff, Santiago Lombeyda, Maggie Hendrie, Hillary Mushkin
This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, and was sponsored by the JPL / Caltech / Art Center Data to Discovery Program, and the National Aeronautics and Space Administration (80NM0018D0004).
Please view the LICENSE.md file in the repository for more information on how VECTOR is licensed.