Recognition of pathway directions based on nonlinear least squares method

Abstract

In-pipe inspection is critical for using a pipeline safely. In recent times, the efficiency of inspection via the use of in-pipe robots has been emphasized in the literature. Inspection efficiency can be significantly improved if the in-pipe robot can move autonomously in the pipe. To do this, recognizing the pathway in front of the robot is a key factor. One method to recognize the pathway using the combination of a laser spot array (LSA) and a camera has been proposed long time ago. However, in this paper, the pathway direction is calculated with different way. The most different point is that the central vector of the pipe (the pathway direction) is estimated by Nonlinear least squares method with a constraint on the pipe radius. At last, their performances between our method and the existing method is compared in terms of accuracy and computation time.