Normalize Image

This function block rescales image pixel values using selectable normalization methods to improve visualization or adjust dynamic range. It also supports clipping based on image statistics to reduce the effect of outliers.

📥 Inputs

Reference The primary image to be normalized. (input socket)

Second Image Optional image used to compute normalization statistics. If provided, its statistics are used instead of the reference image. (input socket)

📤 Outputs

Image Any The normalized image ready for further processing or display. (output socket)

🕹️ Controls

Morph Type Choose the normalization method. Common options include:

• NORM_MINMAX Rescales pixel values to a specified range for visualization.

• NORM_INF Bounds values relative to the maximum value, useful for dynamic range adjustments.

Pixel Range Select the target output range for pixel values using a horizontal range slider. This controls the minimum and maximum values after normalization.

Clip Intensity Adjust how aggressively pixel values are clipped around the mean (expressed as a percentage). Higher values reduce extremes by clipping values outside mean ± (clip × std).

⚙️ Running mechanism

• When executed, the block reads the selected Morph Type, Pixel Range, and Clip Intensity controls and applies normalization to the Reference image.

• If Second Image is connected, its mean and standard deviation are used to determine clipping ranges; otherwise the statistics of the Reference image are used.

• The block first clips extreme values around the mean according to Clip Intensity, then maps pixel values into the chosen Pixel Range using the selected normalization method, producing the output image.

🎨 Features

• Quick selection between visualization-friendly scaling (NORM_MINMAX) and maximum-bound scaling (NORM_INF).

• Adjustable clipping to reduce the effect of outliers or noisy pixels.

• Option to use a separate image for normalization statistics, enabling consistent normalization across multiple images or camera streams.

📝 Usage instructions

1. Connect the image to be normalized to the Reference input.

2. Optionally connect a stable image or background sample to the Second Image input to use its statistics.

3. Select Morph Type appropriate for your goal:

   • Use NORM_MINMAX to set explicit display ranges.

   • Use NORM_INF when you want to bound by maximum values.

4. Adjust Pixel Range to define the output min/max values.

5. Tune Clip Intensity to limit outlier influence before normalization.

6. Feed the resulting Image Any into visualization or downstream processing blocks.

💡 Tips and Tricks

• If your input images are very large, add Image Resizer before this block to speed up preview workflows.

• Use Denoising or Blur before normalization to reduce noise-driven clipping and get more stable statistics.

• For histogram-driven contrast tuning, try Auto Contrast or Contrast Optimization after normalization to refine visual appearance.

• To inspect results visually, connect the output to Show Image or save examples with Image Logger / Image Write.

• When preparing images for detection tasks, normalize consistently across frames by supplying a representative image to the Second Image input (for example, a background sample) so detections downstream (like Find Object or Object Detection) are less sensitive to lighting changes.

🛠️ Troubleshooting

• If output looks washed out or clipped, reduce Clip Intensity or widen the Pixel Range.

• If normalization appears inconsistent across frames, provide a stable image to Second Image so the same statistical reference is used every run.

• If no image appears, ensure a valid image is connected to Reference and preview with Show Image.