eroniki / build_daq Goto Github PK

By using features and the locations, match people between two frames taken at different timestamps.

By using a framework like Bootstrap, the user interface should be designed to create a more pleasing look for the project.

Need to find a way to store extrinsic parameters.

• Define the events to be detected
• Develop algorithms to detect each event individually
• Package and automate each detection algorithm as an analysis tool

Is your feature request related to a problem? Please describe.
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]

Describe the solution you'd like
A clear and concise description of what you want to happen.

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
Add any other context or screenshots about the feature request here.

For each detected person in each frame:

• create a bounding box (a thumbnail image)
• save the bounding box image into its corresponding location
• extract features on the bounding box by using the vision package
• save the feature locations and descriptors as numpy files by using the utils package

Is your feature request related to a problem? Please describe.
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]

Describe the solution you'd like
A clear and concise description of what you want to happen.

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
Add any other context or screenshots about the feature request here.

The method matching descriptors has a weird name for some reason.

• Visualization tool for 2d detections
• Visualization tool for 3d estimations
• Visualization for matching algorithm
• Video exporting from the different visualization tools

By using the extracted features, match people accross cameras:

• Use vision package to match features

There are some many coordinate frames defined in the project (one for each camera) some of which are related to each other with extrinsic parameters.

• Read and parse devices.json to create a tree-like structure to define camera relations
• By using standard libraries, implement transformation between cameras (we can depend on this for the most part.)
• Define a global coordinate frame by using one of the cameras
• Project all the joints from stereo camera coordinate frame to the global

By using the computer_vision package, create an estimate for each matched persons.

• Create the folder structure for each frame each camera and each experiment.
• For each frame, save the detected people into a json file to its corresponding location by using the utils package.