STCI

STCI stands for SpaceTelescopeColorImage.

STCI is a small astronomy image-composition tool for making display-ready color JPEGs or 16-bit TIFFs from space telescope cutouts and other aligned mono images.

Installation

pip install STCI

The installed Python import module is STCI.

Example usage

mk_colorimg creates one color image from either a 3-channel RGB array or three mono images in (R, G, B) order. For Euclid-style color images, the default mapping is NIR_J, NIR_Y, VIS.

from STCI import mk_colorimg

mk_colorimg(
    [
        "cutout_H.fits",    # R channel
        "cutout_VIS.fits",  # G channel
        "cutout_Y.fits",    # B channel
    ],
    output_jpg="target_mtf_vis_y_h.jpg",
    input_mode="raw",
    ReplaceL=False,
    reference_ROI=200,
)

For a NumPy RGB cube:

from STCI import mk_colorimg

mk_colorimg(rgb_array, output_jpg="target_color.jpg", input_mode="normalized")

Use a .tif or .tiff output filename to write a 16-bit TIFF instead of an 8-bit JPEG:

mk_colorimg(rgb_array, output_jpg="target_color.tiff", input_mode="normalized")

Set ReplaceL=False to skip replacing the CIELab L* channel with the stretched blue luminosity channel. By default, mk_colorimg keeps the original replacement behavior.

Set reference_ROI=200 to estimate the raw normalization, color calibration, and STF/HT stretch from the centered 200 x 200 pixel region, then apply those settings to the full image. By default, reference_ROI=None estimates from the full image.

For raw three-channel inputs, STCI uses one shared linear scale derived from the largest channel 99th percentile. This preserves relative RGB amplitudes while preventing a small number of saturated stars from making the target galaxy too dark or excessively warm.

Single-Band MTF Image

mk_monoimg creates one grayscale image from a single mono image. It uses the same MTF-style stretch as mk_colorimg, but omits RGB-only steps such as color calibration, Lab luminance replacement, SCNR, and saturation.

from STCI import mk_monoimg

mk_monoimg(
    "cutout_VIS.fits",
    output_jpg="target_vis_mtf.jpg",
    input_mode="raw",
)

For a NumPy mono image:

from STCI import mk_monoimg

mk_monoimg(mono_array, output_jpg="target_mono.jpg", input_mode="normalized")

Download a Euclid Color Image

Euclidimg downloads Euclid DR1 VIS, NIR_Y, NIR_J, and NIR_H FITS cutouts, then renders one color image using the NIR_J / NIR_Y / VIS channel order.

from STCI import Euclidimg

result = Euclidimg(
    ra=50.7163333,
    dec=-39.7693889,
    size=5.0,
    path="euclid_color",
    cred="Euclid/cred.txt",
    output_jpg="EUCLJ032251.92-394609.8.jpg",
    ReplaceL=True,
    RGB="auto",
    reference_ROI=None,
    style="Euclid",
)

print(result["jpg"])
print(result["fits"])
print(result["rgb_bands"])

Arguments:

• ra, dec: target coordinates in degrees.
• size: cutout radius in arcsec. For example, size=5.0 makes a 10" x 10" image.
• path: output directory for the FITS files and rendered image.
• cred: Euclid credentials file passed to astroquery.esa.euclid.
• output_jpg: optional output filename written inside path; use .tif or .tiff for 16-bit TIFF.
• ReplaceL: if False, skip replacing the CIELab L* channel with the stretched blue luminosity channel.
• RGB: Euclid bands in (R, G, B) order. Use "auto" for the default available-band choice, or pass bands such as ("NIR_H", "NIR_Y", "VIS") or ("H", "Y", "VIS").
• reference_ROI: optional centered square ROI size, in pixels, used to estimate the display scaling before applying it to the full image.
• style: rendering workflow. "Euclid" is the unchanged original STCI workflow. "Euclid Epic" uses the H/Y/VIS workflow described below.

The returned dictionary contains the selected FITS paths and the final rendered image path. If the first overlapping mosaic tile for a band is empty or all zero, the downloader tries the next matching tile.

Euclid Epic style

style="Euclid Epic" selects NIR_H / NIR_Y / VIS and applies the display workflow validated on the Euclid--SGA image set:

• linked-STF target_background=0.125
• fixed rank-0032 H/Y/VIS color-calibration gains
• ChannelCombination R=H, G=Y, B=VIS; this is the default Epic mapping. The earlier equal mix remains available by explicitly setting EuclidEpicSettings(green_vis_weight=0.5)
• a VIS-H > 1.25 mag red-source mask with VIS-only point-source protection; H and Y are never used to classify point sources, and the Euclid VIS point spread function is treated as having a typical FWHM of 1.6 pixels. The existing compact-peak detector is augmented with DAOStarFinder detections whose native VIS peak is above the sigma-clipped background by 5 sigma and whose mean positive brightness within 3 pixels is at least 10 times the mean in the 3--6 pixel annulus; DAO and existing detections are merged within a 2-pixel matching radius
• a bright-VIS host veto, validated on rank 1539, that measures VIS light on a 12-pixel scale and rejects red-mask ellipses inside connected host regions brighter than 5 times the native-pixel VIS noise and at least 500 pixels in area; this prevents dust in luminous galaxies from receiving the red-source Screen enhancement
• one equivalent feathered ellipse per accepted red source
• a post-STF H-to-VIS Screen luminance blend with strength 0.58
• the original STCI post-HT, SCNR, and saturation chain, followed by the Euclid--SGA final SCNR pass

from STCI import Euclidimg

result = Euclidimg(
    ra=39.446736,
    dec=-33.822218,
    size=52.965,
    path="PGC671065_epic",
    cred="Euclid/cred.txt",
    output_jpg="PGC671065_EuclidEpic.tiff",
    style="Euclid Epic",
)

The Epic style is a display composite and must not be interpreted photometrically. The style changes rendering only; Euclidimg retains the existing Euclid FITS downloader. To render already restored BGSUB+BGMOD inputs, call mk_euclid_epicimg([H_fits, Y_fits, VIS_fits], output_jpg=...) directly.

Download Euclid FITS Only

from Download_Euclid import EUC_download

fits_paths = EUC_download(
    ra=50.7163333,
    dec=-39.7693889,
    size=5.0,
    path="euclid_fits",
    cred="Euclid/cred.txt",
)