Detect Reference

This function block is used to automatically identify objects in an input image based on a reference image. It supports various options for fine-tuning the detection process.

📥 Inputs

Image The main image in which the reference object needs to be detected.

Reference The reference image of the object that should be matched.

Mask An optional mask image that can isolate areas to consider during the detection process.

📤 Outputs

Image Any The output image showing the detected objects with matching regions highlighted.

Rectangle Coordinates The coordinates of rectangles that bound the detected references, supporting multiple detections.

Match Percentage The percentage of match for each detected object.

🕹️ Controls

Match Threshold % A slider that sets the threshold for acceptance of matches. By tweaking this value, you can minimize false positives.

Down-size A slider that reduces the size of the reference image to improve processing speed, allowing for faster detection.

Rotations A slider to set the number of rotation slices for detection of objects in various orientations.

Sweep Angle A slider range that defines the angle sweep for rotations during object detection.

Horizontal Flip Search A checkbox that includes consideration for flipped versions of the reference image, which can aid in detection under certain conditions.

Estimation Method A dropdown menu to select the method of estimation for object matching.

Color Mode A dropdown menu to select whether to process images in grayscale or color (BGR), with color processing typically increasing accuracy.

🎨 Features

Rotational Detection Effectively detects objects from multiple rotational angles, enhancing detection capabilities.

Mask Support The option to use a mask helps concentrate detection on vital areas of the input image, improving efficiency.

Visual Feedback The resulting image output clearly displays matched areas, providing real-time feedback on detection results.

📝 Usage Instructions

1. Connect Input Images: Link the input image to the Image input, and the reference image to the Reference input. Optionally, connect a mask to the Mask input.

2. Set Parameters: Use the sliders and dropdowns to define the desired match threshold, size reduction, rotations, and color mode.

3. Evaluate: Run the block to perform object detection. The resulting output will display the matches found in the image.

📊 Evaluation

When executed, this function block effectively analyzes the input image based on the reference provided and outputs the matched image, coordinates, and match percentages.

💡 Tips and Tricks

Improving Detection

When using the detection, ensure that the reference image is clear and well-lit. Adjusting the Match Threshold % can also help in increasing the accuracy of detections.

Using Masks

If only specific parts of the input image are of interest, utilize the Mask input to limit the areas considered by the detection process, enhancing both speed and accuracy.

Rotation Adjustments

Experiment with the Rotations and Sweep Angle controls to find optimal settings for detecting objects that may appear in various orientations.

🛠️ Troubleshooting

No Matches Found

If you receive no matches, verify that the reference image accurately represents the object in the input image and adjust the Match Threshold % to a lower value for a more lenient match.

Performance Issues

If performance is lagging, consider reducing the Down-size setting, which will allow the detection process to operate on a smaller image, speeding up computation.

None Detected Areas

If certain regions of the input image do not yield results, check the Mask input for any potential blockages or enhance the reference image against the specific instances you expect to detect.