Corner Detector

This function block detects corner patterns in an image and highlights them for visual inspection. It is useful for feature extraction, alignment checks, and preparation for shape or tracking tasks.

📥 Inputs

Image Any This is the image to analyze. Accepts grayscale or color images.

📤 Outputs

Detected Corners An image with detected corners drawn for quick visual feedback.

Number of Corners Total count of detected corner points.

Locations List of corner coordinates (multiple points) for downstream processing.

🕹️ Controls

Detector Type Choose the corner detection method (options include HARRIS CORNER and SHI-THOMAS CORNER).

Threshold Adjusts sensitivity for how many corners are detected. Higher values reduce detections; lower values increase detections.

🎨 Features

• Visual markers drawn on the input image to show each detected corner.

• Option to switch between two widely used corner detection algorithms.

• Returns both visual output and structured data (count and coordinates) for further processing.

⚙️ Running Mechanism

When provided an image through the input socket, the block analyzes local intensity patterns to locate corner points. It marks each detected corner on a copy of the input image and outputs the annotated image, the number of corners found, and the list of their coordinates. You can switch detection method and tune the Threshold to control sensitivity.

📝 Usage Instructions

1. Connect an image source to the Image Any input.

2. Select the desired Detector Type.

3. Adjust Threshold until the number of detected corners matches your expectation.

4. Use the Detected Corners output to preview results and Locations output for further automation.

💡 Tips and Tricks

• Preprocess noisy images with Blur or Denoising to remove spurious corners from texture noise.

• Improve contrast and visibility of features with Contrast Optimization or Adjust Colors before detection.

• Limit the search area using Image ROI Select or Image ROI so only the region of interest is processed.

• If the image is very large and detection is slow, use Image Resize to downscale before running this block.

• Visualize results larger or inspect details using the Show Image block.

• Combine corner outputs with Find Contour or Feature Detector when you need higher-level object localization or matching.

• Store or freeze a good frame for reference with Image Memory or save examples via Image Logger.

🛠️ Troubleshooting

• If too many corners are detected, raise the Threshold or apply smoothing with Blur.

• If too few or no corners are detected, lower the Threshold or increase contrast with Contrast Optimization.

• If detected points appear outside the area of interest, crop the input with Image ROI Select before feeding this block.

• For inconsistent results across lighting conditions, try Adjust Colors or disable automatic camera adjustments upstream (use camera advanced settings if available).