Dynamic Edge Threshold for Marker Detection Accuracy

The patent applies dynamics by replacing the fixed threshold with a dynamic threshold that adapts to varying lighting conditions. The threshold is calculated based on the luminance values of pixels in the captured image, allowing the binarization process to adjust automatically to different brightness levels, thereby maintaining detection accuracy across various lighting environments while keeping the processing relatively efficient.

The patent changes the parameter of the threshold from a constant value to a variable value derived from image data. By calculating the threshold based on actual pixel luminance values in the captured image, the system adapts to different lighting conditions, resolving the contradiction between simple processing and accurate detection in varying environments.

The patent replaces conventional mechanical edge detection algorithms with a physics-based approach using luminance thresholding. By substituting complex mechanical processing with a simpler optical principle (comparing pixel luminance values against a calculated threshold), the system achieves both reduced processing load and improved accuracy in challenging conditions like partial occlusion and reflection.

The threshold calculation process uses the image data itself to determine the appropriate threshold value, making the system self-adapting. The luminance values from the captured image are used to compute the threshold, which then automatically adjusts to handle various lighting and occlusion scenarios without requiring complex external processing.