Automatic driving processing method and apparatus based on scene segmentation, and computing equipment

Abstract

The invention discloses an automatic driving processing method and apparatus based on scene segmentation, and computing equipment. The automatic driving processing method based on scene segmentation includes the steps: acquiring the image of the current frame in the video shot and / or recorded by an image acquisition device in real time; inputting the image of the current frame into a scene segmentation network to obtain a scene segmentation result corresponding to the image of the current frame; according to the scene segmentation result corresponding to the image of the current frame, determining the contour information of a specific object; according to the contour information of the specific object, computing the opposite positional relation between the self vehicle and the specific object; and according to the opposite positional relation which is obtained through computation, performing automatic driving control on the self vehicle. The technical scheme of the automatic driving processing method and apparatus based on scene segmentation can accurately obtain the scene segmentation result corresponding to the image of the frame in real time, can more accurately computing the opposite positional relation between the self vehicle and the specific object on the basis of the scene segmentation result, can perform automatic driving control according to the opposite positional relation which is obtained through computation, and is conductive to improving the safety of automatic driving.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a scene segmentation-based automatic driving processing method, device, computing device and computer storage medium. Background technique [0002] In the existing technology, image scene segmentation processing methods are mainly based on fully convolutional neural networks in deep learning. These processing methods use the idea of ​​transfer learning to migrate the network obtained through pre-training on large-scale classification data sets to image processing methods. The segmentation network is trained on the segmentation data set to obtain the segmentation network for scene segmentation, and then the segmentation network is used to segment the scene of the image. [0003] The network architecture used in the segmentation network obtained in the prior art directly uses the image classification network, and the size of the convolution block in the convolution...

AI Technical Summary

Problems solved by technology

However, for image scene segmentation, the scene often contains objects of different sizes. Using a segmentation network with a fixed-size receptive field often causes problems when dealing with too large and too small objects. For example, for smaller objects, the receptive field It will capture too much background around the target, thereby confusing the target with the background, causing the target to be missed and misjudged as the background; for larger targets, the receptive field can only capture a part of the target, which makes the target category judgment biased, resulting in Discontinuous Segmentation Results

Therefore, the image scene segmentation processing method in the prior art has the problem of low accuracy of image scene segmentation, and the segmentation result obtained by using the existing image scene segmentation processing method in the automatic driving scene cannot calculate itself well. The relative positional relationship between the vehicle and other specific objects such as other vehicles, pedestrians, and roads in the frame image of the collected video, so that the safety of automatic driving cannot be guaranteed

Images