Find Reference

This function block is designed to locate a specific object within an input image based on a provided reference image. It employs feature detection and matching techniques to accomplish this task.

📥 Inputs

Object Image An image of the object you wish to locate within the reference image.

Reference Image A template image that serves as the reference for locating the object.

📤 Outputs

Result Image The modified image with the detected object highlighted by bounding boxes.

Detected Object Image The corrected image of the detected object after applying perspective transformations.

Bounding Box Coordinates of the bounding box that encloses the detected object.

Corner Coordinates The coordinates of the corners of the detected object's bounding box, which can include multiple detections.

Center Position The coordinates of the center position of the detected object.

🕹️ Controls

There are no specific controls in this block, as it automatically evaluates upon receiving the necessary input images.

🎨 Features

Robust Feature Detection Utilizes ORB feature detection to identify and match keypoints between the object and reference images.

Visual Highlighting Outputs an image that clearly marks the detected object and its bounding box.

Perspective Transformation Provides the option to obtain a corrected view of the detected object, useful for further processing or analysis.

📝 Usage Instructions

1. Input Images: Connect the object image to the Object Image input and the reference image to the Reference Image input.

2. Evaluate: Run the block to process the images. The output will include the result image, detected object image, bounding box, corner coordinates, and center position.

📊 Evaluation

When executed, this function block outputs the necessary images and coordinates that describe the detected object based on the reference image, including any relevant transformation to align the object correctly.

💡 Tips and Tricks

Enhancing Detection

Ensure that the reference image is clear and resembles the object in various scales and rotations to improve detection accuracy. Preprocessing with Blur, Image Threshold, or Auto Contrast can help in achieving better results.

Adjusting Reference Image

If the object is not detected, consider adjusting the reference image scale, brightness, or orientation to better match how the object appears in the object image.

Tracking Multiple Objects

To track multiple objects simultaneously, consider using separate instances of the Find Reference block with different combinations of reference and object images.

🛠️ Troubleshooting

No Object Detected

If the output indicates that no object was found, check the following:

• Ensure both images are loaded correctly.

• Verify that the reference image is a clear representation of the object in question.

• Consider adjusting the scale or rotation of the reference image.

Error Messages during Processing

In case of exceptions or errors during processing, ensure that both the object and reference images are valid and in the correct format. Any corruption or issue with image loading can lead to errors.