Dehaze is an image processing adjustment that estimates how much atmospheric haze, fog or veiling light covers each part of a photograph and then removes it, or adds it, to change contrast and color saturation. In photography it is offered as a single slider in raw converters such as Adobe Lightroom and as similar tools in Capture One and other editors.
Unlike global contrast or a tone curve, dehaze is spatially adaptive: it treats distant, hazy regions more strongly than nearby, clear regions. Within image processing, it sits between global tonal controls and local contrast tools such as clarity, and it is one of the most powerful and most easily overused sliders in a modern editor.
How dehaze models atmospheric haze
Dehaze algorithms are based on an atmospheric scattering model. In this model, each pixel is a mix of the light reflected by the subject, weakened by the air between subject and camera, and airlight, the bright, low-contrast light scattered toward the camera by particles and water droplets. The farther the subject, the lower the transmission and the larger the share of airlight.
To remove haze, the software estimates two unknowns: the color of the airlight and a transmission map that says how much haze covers each pixel. A widely known approach, the dark channel prior, relies on the observation that clear outdoor regions usually contain some very dark pixels in at least one color channel; where no dark pixels exist, haze is assumed. The software then subtracts the estimated airlight and stretches the remaining signal back to full contrast.
| Component | What it represents | Effect on the result |
|---|---|---|
| Airlight estimate | Color and brightness of the haze | Sets which tint is removed |
| Transmission map | Haze density for each pixel | Decides where correction is strongest |
| Edge-aware refinement | Smoothing that follows object edges | Prevents halos around subjects |
| Contrast restoration | Re-stretching of the remaining signal | Raises contrast and saturation |
Dehaze compared with clarity, contrast and texture
Dehaze, contrast, clarity and texture all increase perceived punch, but they act on different scales and for different reasons. Contrast is a global tone change, clarity boosts midtone local contrast over medium-sized areas, texture works on fine detail, and dehaze acts on large regions according to estimated haze, also shifting color.
| Adjustment | Spatial scale | Affects color | Typical use |
|---|---|---|---|
| Dehaze | Large regions, haze-dependent | Yes, strongly | Distant landscapes, fog, smoke, underwater |
| Contrast | Whole image | Slightly | Overall tonal punch |
| Clarity | Medium local areas | Slightly | Midtone structure, rocks, architecture |
| Texture | Fine detail | Little | Skin, fabric, foliage detail |
| Tone curve | Whole image by tone | Depends on mode | Precise tonal shaping |
Because dehaze also raises saturation and darkens shadows, it often needs a small correction afterward: lifting the shadows, reducing vibrance or adjusting white balance. Treating it as a pure contrast slider is a common reason for heavy, oversaturated results.
Positive and negative dehaze values
In most editors the dehaze slider runs from minus 100 to plus 100, with zero meaning no change. Positive values remove haze; negative values add a soft veil of airlight, which lowers contrast and creates a misty or dreamy look. The useful range depends heavily on the image, and large values rarely look natural.
| Setting | Visible effect | Typical subject |
|---|---|---|
| Minus 30 to minus 10 | Gentle glow, softer contrast | Portraits, romantic or misty moods |
| Plus 5 to plus 20 | Clearer distance, natural color | Landscapes with light haze |
| Plus 20 to plus 40 | Strong clearing, deeper sky | Heavy haze, smoke, underwater scenes |
| Above plus 50 | Dark vignetting, noise, color shifts | Rarely useful except as a mask effect |
Dehaze is also available as a local adjustment in brushes, gradients and masks. A graduated mask over a distant mountain range usually gives a cleaner result than a global setting, because the foreground, which has little haze, stays untouched.
Where dehaze fits in a raw editing workflow
Dehaze works best on raw files after white balance and exposure are set, because the airlight estimate depends on overall color and brightness. It is normally applied before noise reduction and sharpening decisions are final, since strong dehazing amplifies noise and can make sharpening halos more visible.
- Set white balance and exposure first, then dehaze.
- Use a moderate global value and refine distant areas with a mask.
- Check shadows and sky edges at 100 percent for noise and halos.
- Recheck saturation and vibrance after dehazing, since color often needs to be reduced.
Limitations and common mistakes
Dehaze cannot recover detail that the haze has completely hidden; it only re-stretches the signal that remains, so it also stretches noise. In deep haze or fog, a large value exposes banding, chromatic noise and blotchy skies, especially in 8-bit JPEG files.
| Problem | Cause | Fix |
|---|---|---|
| Dark, heavy corners | Algorithm reads vignetting as haze | Correct vignetting first, reduce amount |
| Blue or cyan sky cast | Airlight color misjudged | Adjust white balance or tint after dehazing |
| Halos around skylines | Transmission edges not fully refined | Use a masked, lower setting |
| Noisy shadows | Signal stretched in dark areas | Raise shadows, apply noise reduction |
Another mistake is using dehaze on scenes where the haze is the atmosphere the photographer wanted. Morning mist in a valley often separates layers of hills, and removing it entirely destroys depth and aerial perspective.
In simple terms
Haze is like a thin, milky curtain between the camera and distant objects. The farther away a hill is, the thicker that curtain looks. Colors fade and contrast drops.
The dehaze slider estimates how thick the curtain is in each part of the photo. Then it pulls it back, more in the distance and less up close. Moving the slider the other way does the opposite and adds mist. The key is restraint: a small amount looks natural, while a large amount looks dark and artificial.
A practical example
A photographer hikes to a viewpoint on a warm summer afternoon and photographs a chain of mountain ridges about 20 to 40 kilometers away. The raw file looks flat: the far ridges are pale blue-gray and barely separate from the sky. A global dehaze of plus 15 improves the whole frame, but the foreground grass turns too dark and saturated.
She resets the global value to plus 5 and adds a graduated mask with dehaze at plus 30 over the distant ridges only. The far mountains regain their shape and color, while the nearby meadow stays natural. She then raises the shadows slightly and reduces vibrance by a few points. The final image shows clear distant ridges with a believable sense of depth, instead of an overprocessed, high-contrast look.