Road area detection method

Abstract

The invention discloses a road area detection method which comprises the following steps: S1, designing a road enhancement data enhancement method, and generating a road data enhancement function; S2,adopting a road data enhancement function, inputting a training sample, and outputting enhanced road data; S3, designing and training a feature coding network model, and outputting a coding feature map through the feature coding network model by using the enhanced road data; and S4, performing designing and training to generate a road segmentation decoding module and a road type classification decoding module, adopting the road segmentation decoding module and the road type classification decoding module, and inputting the coding feature map to output a road segmentation result and a road type classification result. The road area pixel-level segmentation and multi-type classification results are provided, which can be used for detecting passable areas of intelligent vehicles, and providesa basis for obstacle avoidance and path planning of the intelligent vehicles.

Description

technical field [0001] The invention relates to the field of automatic driving, in particular to a road area detection method based on geometric edge segmentation optimization and deep learning feature encoding and decoding. Background technique [0002] With the rapid development of deep learning technology and intelligent vehicle technology, deep learning has been widely used in environment perception and decision planning in the field of intelligent vehicles. As the core component of the intelligent vehicle environment perception system, road area detection is the basis for intelligent vehicle obstacle avoidance and planning. If camera sensor data is used as input, for each pixel in the image channel, the goal of road area detection is to determine whether the pixel category belongs to the road area or non-road area (obstacle area). Compared with other sensors, monocular cameras have received extensive attention and been deeply researched due to their rich semantic featu...

AI Technical Summary

Problems solved by technology

Although these DCNN-based methods have better segmentation performance than traditional methods, there are still problems such as rough road edge segmentation, sensitivity to road types, and limited segmentation accuracy.

[0004] Generally speaking, the road area detection for intelligent vehicles at this stage has the following problems: 1) it cannot be generally applied to complex urban road types, including urban lane-free lines, urban single-lane and urban multi-lane; 2) there is no Fully consider the characteristics of pixels in different road areas, and do not distinguish between road edge and road center pixels, resulting in poor segmentation results at road boundaries; 3) In common public datasets, the number of training samples with different road geometries is unbalanced, making it vulnerable to Interference such as road types with different road geometries, resulting in insufficient robustness of road segmentation

Images