A road marking detection method based on three-dimensional measurement of line structured light

Abstract

The invention discloses a road index line detection method based on linear structured light three-dimensional measurement. The method comprises eight steps of obtaining pavement gray-scale data, performing gray-scale correction on the obtained pavement gray-scale data, performing smooth filtering processing on the data subjected to the gray-scale correction, performing step edge point detection on the data subjected to the smooth filtering processing, performing index line edge point screening, connecting interrupted edges to form a complete edge, removing a false index line edge, extracting index line information and the like. The method can effectively guarantee completeness of index line detection.

Description

technical field [0001] The invention belongs to the technical field of image recognition, and relates to a road marking detection method, in particular to a road marking detection method based on three-dimensional line structured light measurement. Background technique [0002] Road traffic markings are mainly sprayed with paint (or thermoplastic materials, road buttons, etc.) or installed on the road surface or the vertical surface of obstacles to warn vehicles and pedestrians, express instructions, warnings, prohibitions and guide the way. content in order to guide the driver to ensure driving safety. At the same time, road traffic markings are also an important sign for active positioning and automatic navigation of smart cars. Since road markings will be worn, polluted, and peeled off after a period of use, regular inspection and maintenance of road traffic markings is required. At present, the collection and evaluation of highway marking maintenance data mainly adopts...

AI Technical Summary

Problems solved by technology

[0013] However, due to the uneven gray distribution of the pavement image, the pavement markings may have different gray mean values ​​at different positions in the entire pavement image; due to different shooting environments, the location and area of ​​the markings in the image The size is different (such as the sidewalk area and the non-sidewalk area). In addition, the grayscale features of the worn, polluted, and peeled areas of the marking surface are not significant. Therefore, methods based on threshold segmentation often have over-segmentation and under-segmentation.

Images