Method for improving instance segmentation based on unmanned driving technology

Abstract

The invention discloses a method for improving instance segmentation based on an unmanned driving technology. According to the method, targets are detected and classified on the basis of the Faster R-CNN based on the MASK R-CNN, and then instance segmentation is achieved through FCN feature coarse extraction and CRF optimization output. The method comprises the following specific implementation steps: step 1, classifying targets by using a partial supervision method; step 2, adopting depth separable convolution in the semantic segmentation convolution process to obtain features; and step 3, performing feature fusion optimization on the features obtained by the convolutional layer, introducing semantic information in a low layer, and introducing spatial information in a high layer. According to the method, relatively good target detection and classification results are established at relatively low cost. The method adopts depth separable convolution, thus improving the precision of a segmentation result and the efficiency of a computer, and reducing time loss.

Description

technical field [0001] The invention belongs to a typical problem of computer vision, image recognition and image classification. In particular, the present invention intends to make some improvements to instance segmentation based on unmanned driving technology. The present invention is based on data classification and a method of improving the structure of FCN. [0002] technical background [0003] In recent years, unmanned driving technology has developed rapidly. Unmanned vehicles are intelligent vehicles that sense the road environment through on-board sensing systems, automatically plan driving routes, and control vehicles to reach predetermined targets. In unmanned driving technology, how to effectively, quickly and accurately identify objects and pedestrians in front of the vehicle is undoubtedly a big obstacle to the development of current unmanned driving technology. If it cannot be accurately identified and reacted in a short time, Then unmanned driving technolog...

AI Technical Summary

Problems solved by technology

However, in deep learning, only single-threaded tasks can be completed in a framework. In recent years, with the development of deep learning and computer vision, it is increasingly required to achieve multi-task integration in deep neural networks, namely target detection, image classification , image segmentation is done through a learning framework, the representative framework is instance segmentation, so we want to improve instance segmentation based on the needs of unmanned driving technology

Images