Measure the length of the spiral object in the photo

Hi,

how can i measure the length of the spiral loop with image processing? The following is my input image.

[Jan]

How do you define "length"? This is not clear at the center.

[Pratheek Ullal]

The length is measured in mm along the inner or outer edge of each loop. the image shows the reference scale to measure the spiral loop.

[Pratheek Ullal]

Lenghth is measured in mm along the inner or outer edge of each loop in the spiral coil.

Answer 1

[Arnav]

Hi @Pratheek Ullal,

To measure the length of the black spiral object, you may use the following image processing pipeline in MATLAB:

The steps can be summarized as:

• Convert to grayscale: Since the image is a dark object against a bright background, we can convert it to grayscale without any meaningful loss of data. This is done using rgb2gray function.
• Increase the contrast of the image: This is done to saturate the lower and higher 10% of intensity. This is done using imadjust function.
• Binarize and negate the image: This step converts the adjusted grayscale image into a binary image using adaptive thresholding. The image is then negated to ensure the object of interest is white against a black background.
• Remove smaller components: Small noise and irrelevant components are removed from the binary image using the bwareaopen function, which retains only the components larger than a specified size.
• Morphological closing: This operation is performed to close small gaps in the binary image, using a disk-shaped structuring element. It helps in connecting disjointed parts of the object.
• Skeletonization: The binary image is reduced to a skeletal form using the bwskel function, which preserves the topology of the object while reducing it to a single-pixel-wide representation.

This can be done in MATLAB using the following code snippet:

% Read the image from the specified path
image = imread('path\to\image.jpeg');

% Convert the image to grayscale
grayImage = rgb2gray(image);

% Increase the contrast of the grayscale image
adjustedGrayImage = imadjust(grayImage);

% Binarize the image using adaptive thresholding and negate it
bwImage = imbinarize(adjustedGrayImage, 'adaptive', 'ForegroundPolarity', 'dark', 'Sensitivity', 0.4);
bwImage = ~bwImage;

% Remove small objects from the binary image
bwImage = bwareaopen(bwImage, 100);

% Perform morphological closing to fill small gaps
bwImage = imclose(bwImage, strel('disk', 5));

% Skeletonize the binary image
skeleton = bwskel(bwImage);

% Prune the skeleton to remove small spurs
prunedSkeleton = bwmorph(skeleton, 'spur', 20);

% Label the connected components in the pruned skeleton
[labeledSkeleton, num] = bwlabel(prunedSkeleton);

% Measure the area of each component
componentLengths = regionprops(labeledSkeleton, 'Area');

% Find the largest component based on area
[~, largestIdx] = max([componentLengths.Area]);

% Extract the largest skeleton component
largestSkeleton = ismember(labeledSkeleton, largestIdx);

You can learn more about the functions used here:

https://www.mathworks.com/help/images/referencelist.html?type=function

The identified arc of the spiral is shown below:

Then, the length of the spiral can be estimated as:

length = scale*sum(largestSkeleton(:))

It should be noted that the length of the spiral may not exactly match the real data as there may be imperfections/artefacts in the image. You may fine-tune this by adjusting the parameters in the pipeline.

See if it helps you.