Author(s)

Hengyu Yin, Hongyang Yu

Corresponding Author

Hengyu Yin

ABSTRACT

When the plastering robot works in rooms, it needs to calculate its own precise position by recognizing the ink line on the ground, so that make the robot achieve the operation accuracy. Due to the special environment of the construction site, the traditional image processing algorithm cannot well filter out the interference of non-ink line on the ground, which makes the final detection result have some errors. Most of the noise in images can be filtered by identifying the color feature of ink lines, which not only improves the recognition accuracy but also declines the time of detection. Then a line detection called cankerworm-crawl algorithm (CCA) is used to identify ink lines segment and using the least squares method to fit the ground ink line in the image, so that make the differences of two walls which plastering by robot is in ±0.5mm.

KEYWORDS

Lines detection; Image processing; Ink line detection; Plastering robot