Watershed Transform Segmentation of TM Grayscale Images Based on Gradient Magnitude with Boundary Marking

This MATLAB implementation performs watershed transform segmentation on TM grayscale images using gradient magnitude analysis and marks the boundaries of segmented regions. The source code can be directly copied and executed in MATLAB. To achieve accurate segmentation of TM grayscale images, we utilize a gradient-magnitude-based watershed transform approach. This method analyzes the gradient information of the image to partition it into distinct regions with clear boundaries, then labels each segmented region accordingly. The algorithm works by first computing image gradients to highlight intensity transitions, then applying morphological operations to prepare the image for watershed transformation. The MATLAB source code below implements this method and can be copied directly into MATLAB for execution: % Assign the TM grayscale image to variable 'image' % Assume the image is already loaded and stored in variable 'image' % Calculate image gradient using imgradient function % This computes the magnitude of gradient at each pixel, highlighting edges gradient = imgradient(image); % Apply thresholding to gradient image using imbinarize % Converts gradient image to binary format for watershed processing binaryImage = imbinarize(gradient); % Perform morphological operation to fill holes using imfill % Ensures continuous regions by filling interior gaps in binary image filledImage = imfill(binaryImage, 'holes'); % Apply watershed transform to segmented image % Partitions image into catchment basins based on gradient topology segmented = watershed(filledImage); % Extract region boundaries using bwperim function % Identifies perimeter pixels of each segmented region boundary = bwperim(segmented); % Display original image and segmentation results figure; subplot(1, 2, 1); imshow(image); title('Original Image'); subplot(1, 2, 2); imshow(label2rgb(segmented)); hold on; contour(boundary, 'r'); title('Segmentation Results'); To implement watershed transform segmentation and boundary marking for TM grayscale images, copy the above source code into MATLAB and execute it following the described steps. The algorithm effectively handles gradient-based segmentation while clearly delineating region boundaries through contour overlay.