工业控制计算机2024,Vol.37Issue(4) :106-108.

基于Canny边缘检测的空频结合图像增强算法

Spatial Frequency Combined Image Enhancement Algorithm Based on Canny Edge Detection

彭兆东 涂琦玉 彭烨超
工业控制计算机2024,Vol.37Issue(4) :106-108.

基于Canny边缘检测的空频结合图像增强算法

Spatial Frequency Combined Image Enhancement Algorithm Based on Canny Edge Detection

彭兆东 1涂琦玉 1彭烨超1
扫码查看

作者信息

  • 1. 上海大学通信与信息工程学院,上海 200444
  • 折叠

摘要

针对多普勒图在雷达信号处理中被广泛应用,但是却面临复杂情况下识别准确率不高的问题,提出了一种基于Canny边缘检测的空频结合图像增强算法.该方法首先利用已有的数据集,对原始图片进行直方图均衡化与同态滤波的空频结合处理;然后进行Canny边缘检测操作,旨在增强谱图信息;最后采用KNN(K-Nearest Neighbors)算法对谱图进行识别分类.实验结果表明,所提出算法可有效地将跳跃、慢跑等 6 种动作的识别准确率提高到 94%,验证了该方法处理图谱并进行分类的可行性.

Abstract

In response to the widespread application of Doppler maps in radar signal processing,but facing the issue of low accuracy in complex situations,this paper proposes a Canny edge detection-based spatial-frequency combined image enhancement algorithm.This method first utilizes an existing dataset to perform spatial-frequency combined processing on the original images using histogram equalization and homomorphic filtering.Then,it applies the Canny edge detection oper-ation aiming to enhance spectral information.Finally,the KNN(K-Nearest Neighbors)algorithm is employed for the recogni-tion and classification of the spectral maps.The experimental results indicate that the proposed algorithm can effectively im-prove the recognition accuracy of six types of actions,such as jumping and running,to 94%.This validates the feasibility of the proposed method for processing spectral maps and performing classification.

关键词

行为识别/图像增强/Canny算法/K最近邻算法

Key words

behavior recognition/image enhancement/Canny algorithm/K-Nearest Neighbor algorithm

引用本文复制引用

出版年

2024
工业控制计算机
中国计算机学会工业控制计算机专业委员会 江苏省计算技术研究所有限责任公司

工业控制计算机

影响因子:0.258
ISSN:1001-182X
参考文献量11
段落导航相关论文