Python module for capturing images from the Lepton over SPI (for example, on a Raspberry PI). Uses the Pylepton library for thermal image capture. Integrates GPS coordinates with corresponding thermal image and CSV logging. Contains scripts to convert CSV data into KML to be used with Google Earth image processing. Also contains GUI for accessing functions.
The thermal imaging module requires cv2 and numpy modules. To install:
sudo apt-get install python-opencv python-numpy
Next, to install into site-packages for distribution through a distutils setup:
cd library
sudo python setup.py install
Requires pyKML and lxml. To install:
sudo apt-get install libxml2-dev libxslt-dev python-dev
sudo apt-get install python-lxml
sudo pip install pykml
Requires PyQt4. To install:
sudo apt-get install python-qt4
To install Raspberry Pi onboot script startup using Cron:
sudo apt-get install gnome-schedule
Capture a single frame
import numpy as np
import cv2
from pylepton.Lepton3 import Lepton3
with Lepton3() as l:
a,_ = l.capture()
cv2.normalize(a, a, 0, 65535, cv2.NORM_MINMAX) # Extend contrast
np.right_shift(a, 8, a) # Fit data into 8 bits
cv2.imwrite("output.jpg", np.uint8(a)) # Write it
Image data from capture() is 12-bit, non-normalized (raw sensor data). Here we contrast extend it since the bandwidth tends to be narrow.
capture() returns a tuple that includes a unique frame ID, as lepton frames can update at ~27 Hz, but only unique ones are returned at ~9 Hz. Currently, this is just a simple sum, but ideally this will turn into a real frame ID from telemetry once this feature is implemented.
Note also that the Lepton contructor can take as an optional argument the SPI device on which to find the Lepton. If in your system that device is /dev/spidev0.1, you can instantiate lepton as such:
with Lepton("/dev/spidev0.1") as l:
This program will constantly capture thermal images and log the current GPS coordinates in a CSV file with the corresponding location where the image was taken. These are two separate programs that run independently. Run GPS.py then sensor.py in two separate terminals. Or run main.py which will run the two scripts in the background.
python main.py
To kill the two background processes, you may need to kill the individual processes. To check background processes:
ps aux | grep python
sudo kill <PID>
The program obtains the timestamp from the GPS module so it does not require internet access to run the logger.
Image output is placed into the photos directory where sessions are separated by timestamp. These two scripts are run on bootup on the Raspberry Pi using Cron.
To edit crontab:
crontab -e
Place these two lines inside to start the script on reboot or power cycle
@reboot python /home/pi/flir_lepton_sensor/scripts/GPS.py &
@reboot python /home/pi/flir_lepton_sensor/scripts/sensor.py &
This program will output any image format that opencv knows about, just specify the output file format extension (e.g. output.jpg or output.png)
To capture a png file named output.png:
python capture.py output.png
To view additional options/settings:
python capture.py --help
Additional settings are:
Usage: pylepton_capture [options] output_file[.format]
Options:
-h, --help show this help message and exit
-f, --flip-vertical flip the output image vertically
Requires python-picamera, a Raspberry PI, and compatible camera.
sudo apt-get install python-picamera
To get a 100% lepton overlay (note camera installation still required):
python overlay -a 255
To view additional options/settings:
python overlay.py --help
Additional settings are:
Usage: pylepton_overlay [options]
Options:
-h, --help show this help message and exit
-f, --flip-vertical flip the output images vertically
-a ALPHA, --alpha=ALPHA set lepton overlay opacity
This script will convert CSV data into KML for usage with Google Earth image processing. Resolution/Placemark density can be set by giving the script an argument. The default resolution is create a Placement for every line in the CSV file (with each line equivalent to 1 sample a second). Ex: Resolution = 10 means one Placemark point for every 10 lines in the CSV file.
To convert CSV data into KML file:
python KML_generator_latest.py [resolution]
Example: To generate a Placemark every 10 seconds
python KML_generator_latest.py 10
This script will open a GUI that has the functionality of all these other scripts. Enables image capture and KML generation.
python GUI_window.py