2D Filter

This function block applies a selectable 2D convolution filter to an input image and returns a grayscale filtered result. It is useful for edge detection, embossing, sharpening and similar image enhancement tasks.

📥 Inputs

Image Any Accepts any image (color or grayscale) to be filtered.

📤 Outputs

Image Gray Returns the filtered image in grayscale.

🕹️ Controls

Filter Type Choose the desired filter from a dropdown. Common choices include Emboss, Contour, Sharpen, Laplacian, and several Sobel/edge variants.

🎨 Features

• Applies a variety of predefined convolution kernels for common 2D filtering tasks.

• Automatically converts color input to grayscale before applying the filter.

• Simple dropdown interface to quickly switch between different filter kernels.

📝 Usage Instructions

1. Connect an image-producing block (camera, file loader or other processing block) to the Image Any input.

2. Select the desired filter using the Filter Type dropdown.

3. The block outputs a single-channel grayscale image on Image Gray, suitable for downstream analysis or visualization.

📊 Evaluation

On evaluation this block converts color images to grayscale if needed, applies the selected 2D convolution filter, and outputs the filtered grayscale image.

💡 Tips and Tricks

• For noisy images, place a Blur block before this block to reduce small artifacts and get cleaner edge responses.

• If you want a binary edge result after filtering, connect the output to an Image Threshold block.

• Use Edge Filter or Sobel Filter blocks when you need specialized edge operators; this block is good for quick experimentation with several different kernels.

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

• When working with regions of interest, crop first using Image ROI, Image ROI Select or Image ROI Polygon and then apply the filter only to the area of interest.

• If you need to align sizes before filtering (large images may be slow), use Image Resize to downscale, filter, then optionally upscale.

Suggested block combinations:

• Camera input workflow: Camera USB → 2D Filter → Show Image

• Preprocessed edge detection: Blur → 2D Filter → Image Threshold

• ROI-based filtering: Image ROI Select → 2D Filter → Find Contour

🛠️ Troubleshooting

• If the output looks identical to the input, ensure the input was not already a single-channel image and that a different Filter Type is selected.

• If output is too noisy, add Blur before this block or try a different filter kernel with stronger smoothing characteristics.

• If output size or format is incompatible with downstream blocks, add Image Resize or Merge Channels as needed.