红外图像特征的三维提取技术

3D feature extraction of infrared image

为了提高红外图像的识别能力,需要对红外图像的三维特征进行有效提取,提出一种基于图形自适应渲染和纹理特征分割的红外图像特征三维提取技术。采用三维红外点扫描方法进行红外图像的纹理渲染和点跟踪识别,在红外透射空间中对采集的红外图像进行纹理分割,在不同红外透射纹理区域中进行图像的动态特征融合处理,采用空间邻域边缘融合方法提取红外图像三维特征的亮度和色差信息,实现红外图像的三维重建和特征提取。仿真结果表明,采用该方法进行红外图像特征三维提取的准确度较高,三维特征重构能力较强,动态特征点的匹配性较好,在红外图像识别中具有很好的应用价值。

In order to improve the recognition ability of infrared image,it is necessary to extract the 3 D feature of infrared image effectively. A 3 D feature extraction technique based on image adaptive rendering and texture feature segmentation is proposed. The texture rendering and point tracking recognition of infrared image are carried out by 3 D infrared point scanning method,and the texture of the collected infrared image is segmented in infrared transmission space. Dynamic feature fusion is carried out in different infrared transmission texture regions. The spatial neighborhood edge fusion method is used to extract the brightness and color difference information of the 3 D features of the infrared image,and the 3 D reconstruction and feature extraction of the infrared image are realized. The simulation results show that the proposed method has higher accuracy,better ability of 3 D feature reconstruction and better matching of dynamic feature points,and it has a good application value in infrared image recognition.