Here's a place to share codes processing James Webb's images
The first code is named count-celestials.py which counts the objects on James Webb's First Deep Field released on July 11, 2022 ( You can download the image file from here )
assume the gray-scale image as a 3D surface, where the height of each pixel is the value of brightness. the algorithm would scan the gray-scaled image from high-bright points to low-bright ones, and count the peaks.
to run the code, change IMAGE_PATH in the file and type
python3 count-celestials.py
there are three parameters to configure:
-
BRIGHTNESS_LOWER_THRESHOLD: a number between 0-1, part of the image with lower brightness than this threshold, would not processed -
NEIGBORHOOD_STEPS: number of pixels in neighborhood -
BRIGHTNESS_TOLERANCE: a value between 0-1, the algorithm would assume the bightness with this margin (v +/-BRIGHTNESS_TOLERANCE%) is equal to v
this code counts around 12K objects in James Webb's First Deep Field Image,
P.S. : this algorithm has still some FALSE positive/negative detection, and also the performance can be improved