A semantic segmentation and attention mechanism-combined lane line detection method and system

Abstract

The invention discloses a semantic segmentation and attention mechanism-combined lane line detection method and system, and relates to the technical field of lane line detection, and the method comprises the following steps: inputting a picture to a backbone network to extract lane line features in the picture, inputting the features extracted by the backbone network into a full connection layer, and carrying out the processing of the full connection layer for obtaining local features of each lane line; inputting the local features of each lane line into an attention mechanism module to obtain global features of each lane line, and superposing the global features with the local features to obtain final lane line parameters; and comparing with a real lane line in an input picture, respectively calculating mean square error loss of the lane line and cross entropy loss of a mask graph, calculating a gradient of a corresponding convolutional layer, updating parameters of the convolutional layer and a full connection layer by using a back propagation algorithm, and continuously iterating until the model converges. According to the method, the convergence speed of the model is increased and the robustness of the model is improved under the condition that the operand is not increased.

Description

technical field [0001] The present invention relates to the technical field of lane line detection, and more specifically relates to a lane line detection method and system combining semantic segmentation and attention mechanism. Background technique [0002] Lane line detection is one of the basic tasks in the field of assisted driving. In the assisted driving system, accurately obtaining the position of the lane line is an important prerequisite for subsequent vehicle route planning and lane departure warning; the assisted driving system captures the current driving vehicle through the vehicle camera. In the front driving area, using digital image processing technology, the specific position of the lane line in the image can be located according to the image clues of the lane line. [0003] The general process of a traditional lane line detection system is camera calibration based on monocular images, image cleaning), feature extraction, lane line model fitting, time domai...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the present invention provides a lane line detection method and system combining semantic segmentation and attention mechanism, to solve the problems raised in the background technology, on the basis of accelerating model convergence speed and improving model robustness, effective Solve the problem of missed detection of lane lines in complex scenes such as blur and occlusion

Images