基于计算机视觉的可见光图像彩色成像研究  

Research on color imaging of visible light image based on computer vision

作　　者：刘倩 陈辉 曾健友 LIU Qian;CHEN Hui;ZENG Jianyou(Wenhua College,Wuhan 430074,China)

机构地区：[1]文华学院,武汉430074

出　　处：《激光杂志》2023年第7期105-109,共5页Laser Journal

基　　金：湖北省自然科学基金(No.2020AA01S167)。

摘　　要：为了提升可见光图像彩色成像的有效性,提出基于计算机视觉的可见光图像彩色成像研究方法。首先利用计算机视觉技术建立摄像机图像空间模型,采集空间物体图像;利用色空间变换对采集的图像开展彩色传递,完成灰度图像到彩色图像的转变,从可见光彩色图像中提取该特征点,并消除稳定性较差的极值点,对图像特征点匹配对可见光彩色图像的重构。实验结果表明,当特征节点数量较高时,设计方法对图像(a)的图像特征点检测用时仅为1.92 s,对图像(b)的图像特征点检测用时仅为1.89 s;验证了该方法的实用性强、成像精度高。In order to improve the effectiveness of visible light image color imaging,a research method of visible light image color imaging based on computer vision is proposed.Firstly,the computer vision technology is used to establish the camera image space model and collect the space object image;The color space transformation is used to carry out color transfer on the collected image,complete the transformation from gray image to color image,extract the feature point from the visible light color image,eliminate the extreme point with poor stability,and match the image feature points to reconstruct the visible light color image.The experimental results show that when the number of feature nodes is high,the detection time for image feature points of image(a)and image(b)is only 1.92 s and 1.89 s respectively;The method is proved to be practical and accurate.

关 键 词：计算机视觉 可见光图像 彩色成像 特征点匹配 

分 类 号：TN212[电子电信—物理电子学]