一种基于压缩感知的人群目标密集度检测方法及装置
This paper proposes a crowd density detection method based on compressed sensing to address the problem of inaccurate crowd density detection in occluded environments. By utilizing the measurement domain features of image patches and a Transformer coding network, the accuracy of crowd density detection is improved.
CN118279821BActive Publication Date: 2026-07-17SUZHOU UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SUZHOU UNIV
- Filing Date
- 2024-03-29
- Publication Date
- 2026-07-17
Smart Images
Figure CN118279821B_ABST
Abstract
本发明涉及一种基于压缩感知的人群目标密集度检测方法及装置，属于图像检测技术领域。对待检测图像和待检测图像的背景图像进行划分，得到M个大小为N*N的待检测图像块和M个大小为N*N的背景图像块，计算每个待检测图像块的测量值的归一化均方误差；对测量值的归一化均方误差最大的L个待检测图像块分别进行压缩感知，得到每个待检测图像块的测量域特征向量，将L个待检测图像块的测量域特征向量拼接作为待检测图像的测量域特征向量；将待检测图像的测量域特征向量输入Transformer编码网络，输出待检测图像的目标特征矩阵；将待检测图像的目标特征矩阵输入人群目标感知分类头，输出待检测图像的人群密度分类结果。
Need to check novelty before this filing date? Find Prior Art