maxbbraun / fever Goto Github PK

Hi Max,

Thank you for your grate project! While I was trying to enable camera over UVC, when I do "cmake.." it comes up with this error:

The camera is attached but it says "JPEG not found".

Moreover, can you please explain how you read temperature from the camera? I am asking because I was able to run the camera in another way, but I think I am missing some information in this way.

Look forward to your answer!

Best,

Hi Max,

sry to disturb, but I'd like to better understand which network is implemented below the tflite model. From my understanding, the DetectionEngine only inputs SSD models. Is that the network architecture running below the tflite? Or is it another one (like mobilenet SSD, efficientDet, YOLOVx or else).

Will wait for your. Thank you for the great work you did.

Especially something like detection and tracking could be useful for more accurate readings.

Face detection currently only works in a limited way, because the detector was not made for thermal images. Two possible directions to take from here:

1. Use an RGB camera right next to the thermal camera, run face detection there, and reproject the detections into the thermal camera frame to read the temperature. Could use the Raspberry Pi Camera Module V2 to keep it cheap and simple or a Intel RealSense D400 if #1 requires it.

   • Pro: fairly straightforward implementation
   • Con: additional hardware, reprojection introduces error

2. Make a thermal face detector. This could probably be done by fine-tuning a detector network previously trained on RGB data. This dataset of 1557 thermal images with bounding boxes might be a start. As for the detector, maybe try EfficientDet. In general, TensorFlow Lite compatibility might be preferable so an Edge TPU can be used.

   • Pro: clean all-in-one solution
   • Con: more unknowns in implementation

In either case, #2 may also require face landmarks (eyes, nose, mouth) instead of plain bounding boxes for better accuracy.

See also:

• Projection in RealSense SDK 2.0

Using the Coral Dev Board instead of a Raspberry Pi with a Coral USB Accelerator may give better performance, but it's unclear if the other components will work with it.

Hi Max,
is there any value to obtain the decimal precision for temperature detection?
Thanks in advance

Gianluca

In order to produce a usefully accurate temperature reading, the raw data from the thermal camera should be adjusted using additional data such as the distance of the face, the ambient temperature, and the humidity of the environment. The body temperature is also different from the surface temperature of the skin, which requires additional adjustments.

Ambient temperature and humidity (as well as air pressure and gases) are already available from the BME680 sensor.

Face distance could be determined in one of two ways:

1. Estimate the distance from the size of face detections. Simple bounding boxes might work, but face landmarks (eyes, nose, mouth) may be more accurate.

   • Pro: straightforward implementation, no additional hardware
   • Con: likely not very accurate (assumes uniform face size, sensitive to detection errors)

2. Use a depth camera such as the Intel RealSense D400 series.

   • Pro: higher accuracy
   • Con: expensive additional hardware, reprojection errors

The exact formula to combine these inputs into a temperature remapping function is also TBD.

See also:

• 5 Factors Influencing Radiometric Temperature Measurements
• UAS Radiometric Temperature Measurements

According to maxbbraun/thermal-face the model thermal_face_automl_edge_fast.tflite should run on CPU
However I tried it and it won't work.

Error code is shown below
super().__init__(model_path) File "/usr/lib/python3/dist-packages/edgetpu/basic/basic_engine.py", line 92, in __init__ self._engine = BasicEnginePythonWrapper.CreateFromFile(model_path) RuntimeError: No Edge TPU device detected!

Model I use

flags.DEFINE_string('face_model', 'thermal_face_automl_edge_fast.tflite', 'The TF Lite face detection model file compiled for CPU ' 'CPU.')
Per google documents edgetpu.detection.engine.DetectionEngine can run on CPU .

Is there any problem with thermal_face_automl_edge_fast.tflite ?

how to run thee code without coral ,can you give tutorial bro?