I have a polyline map which I extracted automatically and I want to compare it with another polyline map which I extracted manually in order to evaluate the accuracy of the automatic method. (my study area is not on Earth so there is no ground truth). Generally accuracy assessments is based on a confusion matrix and classes, but I don't have any classes here, they are two series of lines, which I want to see how close is the automatic polyline results to the manual one! I have only position of the lines, I can extract their length, sinuosity, orientation.

Any suggestion for accuracy assessment in this case?


There is a white paper that lays out an approach:

3. Methods for assessing the quality of lines

In the following sections, we will be presenting and discussing methods for calculating and quantifying the geometric accuracy of lines.

Accuracy Assessments of Geographical Line Data Sets, the Case of the Digital Chart of the World*

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.