Feature Detector

This function block locates and identifies objects in an input image based on features extracted from a training image. It utilizes various detection methods to provide robust identification capabilities.

📥 Inputs

Train Image An image containing the object or feature you want to detect in the input image.

Input Image From Camera The real-time image that will be analyzed to detect the features based on the training image.

📤 Outputs

Detected Image The output image that shows the detected features or objects highlighted.

Detect Status A boolean output indicating whether the detection was successful or not.

Center The center point of the detected object, represented by its coordinates.

🕹️ Controls

Homography Type A dropdown menu to select the type of homography method to be used for matching (e.g., RANSAC, LMEDS, RHO).

Compute Type A dropdown menu to choose the computation type for feature matching (e.g., STABLE, PERFORMANCE).

Number of Features A slider to set the number of features to be considered in the matching process.

Distance Threshold A slider to define the distance threshold for matching features, affecting how strictly matches are determined.

K nearest A slider to set the number of nearest neighbors to evaluate during detection.

Pyramid Decimation Ratio A slider to adjust the pyramid decimation ratio for multi-scale feature matching.

Number of Pyramid Levels A slider to set how many pyramid levels to use during detection.

Point Comparison Type A slider to define the comparison method to be used when evaluating features.

🎨 Features

Multiple Detection Algorithms Allows for selection between various homography types and compute methods, providing flexibility based on user needs.

Real-time Detection Analyzes input images in real time, making it suitable for dynamic environments like surveillance or object tracking.

Visual Feedback Provides an output image that visually highlights the detected objects, making results easily interpretable.

📝 Usage Instructions

1. Connect Input Images: Link the training image to the Train Image input and the live image (from a camera or file) to the Input Image From Camera input.

2. Choose Detection Parameters: Adjust the sliders and dropdowns to configure the detection strategy as desired. This includes selecting the Homography Type, Compute Type, and adjusting sliders for feature detection settings.

3. Run the Block: Execute the block to perform the feature detection.

4. Retrieve Results: Check the Detected Image for highlighted detected features, the Detect Status for success confirmation, and the Center output for the detected object's center coordinates.

📊 Evaluation

When executed, this function block will analyze the input image for features that match those in the training image, outputting a modified image, the detection status, and the center of the detected feature.

💡 Tips and Tricks

Use High-Quality Training Image

Ensure the training image is clear and well-lit to improve detection accuracy.

Tuning Parameters

Experiment with the sliders to find the optimal configuration for your specific images and objects. Adjusting the Distance Threshold can significantly affect detection performance.

Monitoring Detection Status

Be sure to monitor the Detect Status output; if detection fails, experiment with your feature parameters and verify the quality of the training image.

Testing on Different Inputs

Test your configuration with various input images to validate robustness. This will help to ensure that your detection setup generalizes well across different scenarios.

🛠️ Troubleshooting

No Detection Occurred

If no features are detected, verify the quality of your training image and check if the parameters, especially the Distance Threshold, are too strict.

Consider adjusting the Number of Features to allow for more flexible matches.

Detection Performance Issues

If performance is slow or inconsistent, try reducing the number of features or adjusting the Compute Type for a balance between accuracy and processing speed.