基于组合式双目相机的测距方法以及计算机可读存储介质
By combining binocular camera calibration and feature matching with the Hungarian algorithm and triangulation method, multi-target ranging was achieved, solving the problems of computational redundancy and high hardware cost in existing technologies, and improving the accuracy and efficiency of ranging.
CN121855458BActive Publication Date: 2026-07-17GUANGZHOU EMBEDDED MASCH TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUANGZHOU EMBEDDED MASCH TECH CO LTD
- Filing Date
- 2025-12-25
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing binocular ranging methods suffer from computational redundancy, high hardware costs, limited applicability, and difficulty in simultaneously measuring the distance information of multiple different targets.
Method used
A combined binocular camera was used for calibration. A pre-trained detection network and feature extraction network were used, and the cosine distance matrix and geometric similarity matrix were combined to perform target matching through the Hungarian algorithm. The distance was calculated using the triangulation method.
Benefits of technology
It supports multi-target ranging, reduces hardware costs, improves matching accuracy, has strong anti-interference capabilities, is easy to calculate and fast, and has a certain degree of scalability.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN121855458B_ABST
Abstract
本发明公开了基于组合式双目相机的测距方法以及计算机可读存储介质，所述测距方法包括以下步骤：对双目相机进行标定和部署，获取相机参数；同步采集图像，调整尺寸后输入预训练检测网络，提取图像信息并处理得到目标图像及目标信息；将目标图像调整尺寸后输入特征提取网络获得特征向量，通过计算特征向量的余弦距离构建余弦距离矩阵，利用目标信息构建几何相似矩阵；通过融合计算两种矩阵后采用匈牙利算法获得匹配目标，在匹配目标区域内搜索最佳匹配点，对最佳匹配点采用三角剖分结合相机参数进行测距。本发明支持多种类多目标同时测距，进一步提升了匹配准确性，不易受到光照和环境变化的影响，并且部署灵活，计算效率高，使用的硬件成本低。
Need to check novelty before this filing date? Find Prior Art