Thermal Inspection
Relations from Other Scenarios
- Digital Reconstruction ← observations ← Thermal Inspection
Relations to Other Scenarios
- Thermal Inspection → visualization → Fire Risk
Summary🔗
This scenario explores how we can observe the thermal state of the construction site.
Models🔗
Intentionally left empty.
Definitions🔗
thermal spot🔗
This is a spot for which we would like to record the temperature.
thermal image🔗
A thermal image is an captured by a thermographical sensor.
The image is stored as a JPEG with the sensor range as additional meta-data in the EXIF (see images (from digital_reconstruction)).
Scenario🔗
As-planned🔗
The user defines the thermal spot as a focus spot (from uxv_recording).
As focus spot (from uxv_recording) is a topic (from topic_management), special information about the thermal spot are indicated as comments (from topic_management).
As-observed🔗
Capturing. The operator (from uxv_recording) takes into account the thermal spots when planning a mission (from uxv_recording).
Apart from specifying the thermal spot as a point of interest (from uxv_recording), the operator (from uxv_recording) should also include thermal cameras in the list of desired sensors (from uxv_recording).
We assume that we use a dual camera that records both the thermal images as well as RGB images (from digital_reconstruction).
The thermal images are stored in images (from digital_reconstruction) .
Inspection. We inspect the thermal images (from thermal_inspection) of the thermal spots:
- Pick a time range,
- Pick a spatial filter (zone, bounding box and other spatial filters, see "Virtual Inspection"); these spatial filters refer to the location where the image was taken and not of where the temperature was observed,
- Pick a temperature threshold (above, below etc.; the complexity of these threshold queries is to be discussed during the implementation).
The relevant thermal images (from thermal_inspection) are presented in a list of triples:
- thermal image,
- the corresponding image (from digital_reconstruction), and
- the location and the view angle where the image has been taken.
(If time permits, we also present an overlay of the thermal image over a RGB image (from digital_reconstruction).)
The user can jump straight to "Virtual Inspection" to inspect the position and the view angle further.
The other aspect sections intentionally left empty.
Analytics🔗
For the filtered images, we provide summary statistics:
- Number of images above threshold
Test Cases🔗
test for magnitude🔗
We automatically generate dummy RGB and thermal images.
They are randomly spread over the time and space.
We make random queries (spatial and temporal) and test that the retrieval time is on the order of 10s of seconds.
Acceptance Criteria🔗
magnitude🔗
The system needs to handle the a lower-digit millions of the thermal images.