电子器件2025，Vol.48Issue(2)：353-358,6.DOI:10.3969/j.issn.1005-9490.2025.02.018
融合机器视觉差学习的无人机自主避障目标检测算法
Autonomous Obstacle Avoidance Target Detection Algorithm for Unmanned Aerial Vehicles Based on Machine Vision Difference Learning
摘要
Abstract
In order to improve the accuracy of drone autonomous obstacle avoidance target detection and efficiently achieve obstacle avoidance,a drone autonomous obstacle avoidance target detection algorithm integrating machine vision difference learning is proposed.The denoising method in machine vision algorithms is used to denoise the flight path data of unmanned aerial vehicles.The improved convolutional neural network combined with Gabor kernel in deep learning algorithm is used for data feature extraction and object detec-tion after denoising.By improving the neural network to calculate the disparity map of image data pairs,introducing Gabor kernels in-stead of traditional convolutional kernels,and combining adaptive optimization algorithms to adjust the weights of network parameters,obstacle targets are accurately located in the disparity map,and drone linear flight path detection is carried out.Finally,the raster scan-ning method is used to determine whether the pixels in the disparity map are obstacle points,achieving object detection and obstacle avoidance decision-making.The experimental test results show that the proposed method can effectively reduce detection time and obtain high-precision autonomous obstacle avoidance target detection results for unmanned aerial vehicles.关键词
机器视觉/改进深度学习/无人机自主避障/目标检测Key words
machine vision/improved deep learning/autonomous obstacle avoidance of unmanned aerial vehicles/object detection分类
信息技术与安全科学引用本文复制引用
陈瑞霞,张云龙,王茜..融合机器视觉差学习的无人机自主避障目标检测算法[J].电子器件,2025,48(2):353-358,6.基金项目
国家自然科学基金项目(62172338) （62172338）
河南省科学技术厅项目(232102210034) （232102210034）
河南省教育厅第九批河南省重点学科(检测技术与自动化装置)建设项目(教高[2018]119号) （检测技术与自动化装置）
