Image data set generation method based on UE5 and AirSim, medium and equipment

Generating simulation scenes through UE5 and AirSim, capturing and analyzing image data, solving the time-consuming and labor-intensive problems in traditional methods, and achieving efficient and flexible image data set generation and model recognition capabilities.

CN120339747APending Publication Date: 2025-07-18SICHUAN XUANJIE INTELLIGENT TECHNOLOGY CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202510394149.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-03-31
Publication Date
2025-07-18

AI Technical Summary

Technical Problem

Traditional image dataset generation methods are time-consuming and labor-intensive, difficult to ensure the consistency and diversity of data, and difficult to meet the needs of large-scale and efficient dataset generation.

Method used

Use UE5 and AirSim to generate simulation scenarios, capture test image data through camera devices, perform model recognition effect analysis, filter out weak links, and generate targeted training image data sets.

Benefits of technology

Reduces costs, improves the flexibility and efficiency of image data set generation, and can generate multiple types of image data sets for specific targets, enhancing the recognition capabilities of the model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120339747A_ABST
    Figure CN120339747A_ABST
Patent Text Reader

Abstract

The invention relates to the field of machine learning training data generation in computer vision, in particular to an image data set generation method based on UE5 and AirSim, a medium and equipment. Comprising the following steps: generating test image data according to UE5 and AirSim; and inputting the test image data into the to-be-tested model, and generating an identification test result corresponding to the test image data. And performing model recognition effect analysis processing on all recognition test results to determine a generation position of training image data corresponding to the to-be-tested model. According to the invention, all objects in a simulation environment are supported by corresponding data parameters in UE5 and AirSim, so that acquired scene data can be converted into corresponding training image data with annotation information. And a large number of multi-type image data sets can be generated. Besides, the training data can be generated in a more targeted manner according to the current weak link of the to-be-tested model, so that the performance of the model can be gradually improved through the training set.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Cited By

  • High-reality simulation data set generation method and system

    CN121904263A