A traffic scene analysis method based on a multi-task network

Abstract

The invention discloses a traffic scene analysis method based on a multi-task network, comprising the following steps: a multi-task network is divided into an encoder, a partition decoder and a detection decoder, wherein the encoder extracts the features of the image and extracts the multi-scale information from the feature map; the segmentation decoder enlarges the size of the feature map and fuses it with the feature map; the detection decoder processes the input characteristic map and outputs a corresponding target detection result; using a deep learning framework Tensorflow to configure, train and test the above multitasking network. The multi-task network of the invention can extract abundant image features, make up for the loss of image detail information caused by downsampling in the encoder, and help to improve the segmentation and detection effect. The invention designs a multi-task network structure, which can realize semantic segmentation and target detection of traffic scene images through one-time back propagation, and has better real-time performance and higher accuracy rate.

Description

technical field [0001] The invention belongs to the field of safety assisted driving, in particular to a multi-task network design method for real-time traffic scene analysis. Background technique [0002] Vision-based traffic scene parsing has important applications in intelligent transportation systems. Semantic segmentation and object detection are two main tasks in traffic scene parsing. Traditional methods to solve these tasks, such as support vector machine (SVM), adaptive boosting algorithm (AdaBoost), random forest iterative method (random forest), etc., have poor generalization and robustness. In recent years, deep learning has made breakthroughs and has been widely used in the field of image processing. The deep learning method automatically extracts rich features from complex data, which makes the model have better generalization ability, and the extracted features can be used for different tasks such as semantic segmentation and object detection. [0003] Sema...

AI Technical Summary

Problems solved by technology

[0004] In the actual traffic scene application, if the traffic scene analysis is carried out and the semantic segmentation and object detection are realized at the same time, two networks must be run at the same time, which requires powerful computing performance and poor real-time performance.

Images