Positioning element data enhancement method and device, equipment and storage medium

Abstract

The invention discloses a positioning element data enhancement method and device, equipment and a storage medium. According to a specific implementation scheme, the method has the following steps: adding noisy points to a to-be-added positioning element to obtain a target positioning element; and processing the target positioning element according to the information of the to-be-processed image, and adding a processing result into the to-be-processed image to obtain a new image. By adoption of the scheme of the invention, the noise points are added to the to-be-added positioning element, so that the image quality of the obtained target positioning element is closer to the image quality after being processed by illumination, corrosion, fading and other reasons in a real environment; meanwhile, the target positioning elements are fused into the to-be-processed image to obtain new image data, so that the quantity of data containing the positioning elements can be increased, the data typedistribution of the positioning elements can be balanced, and the neural network model is ensured to have sufficient quantity and category of training data in the training process. The scheme of the invention can be applied to the field of automatic driving, especially autonomous parking.

Description

technical field [0001] The embodiments of the present application relate to the field of data processing, and in particular, to an enhancement method, device, device, and storage medium for locating element data. Background technique [0002] Deep learning technology has been successfully applied in the visual perception of unmanned driving. Through visual perception, key points on road arrows, signs and some custom signs in natural scenes can be obtained to achieve accurate measurement of vehicle position and attitude. In order to improve the ability and effect of the model in the training process of the neural network model, a large amount of data is often required. These data mainly rely on manual labeling, but only manual labeling data has a long cycle and high cost, and often cannot meet the model's requirements for data quantity and distribution. Require. [0003] Currently commonly used data enhancement methods are: horizontal flipping, cropping, translation, rotatio...

AI Technical Summary

Problems solved by technology

In order to improve the ability and effect of the model in the training process of the neural network model, a large amount of data is often required. These data mainly rely on manual labeling, but only manual labeling data has a long cycle and high cost, and often cannot meet the model's requirements for data quantity and distribution. Require

[0003] Currently commonly used data enhancement methods are: horizontal flipping, cropping, translation, rotation, scaling, affine transformation, color transformation, adding noise, etc. to the data, but these methods can only enhance the original data and cannot enrich the data. category

Images