Two algorithms for line detection, the nonlinear detection and the modified linear detection are examined. Experiments are conducted on detecting suburban roads on Landsat image.
In this paper, we intended to analyze and compare about using the nonlinear detection and modified linear detection algorithm in the two models, 5 by 5 model and 9 by 9 model. And then, we proposed new model which is contained the merit of the two models. We intended to compare between the new model and two models.
It is found that the results will always contain a significant amount of noise. Two processing operators, line connector and line grower are described which reduce the noise by utilizing contextual information of data.