基于SVM模型的单目红外图像深度估计

提出一种通过非线性学习模型来估计单目红外图像深度的算法。该算法首先通过逐步线性回归和独立成分分析(ICA)寻找对于红外图像深度相关性较强的特征,然后以具有核函数的非线性支持向量机(SVM)为模型基础,采用监督学习的方法对红外图像深度特征进行回归分析并训练,在训练过程中通过已知数据回归后的最小均方误差对模型参数进行修正,训练后的模型可对单目红外图像的深度分布进行估计。实验结果证明,利用该模型能较一致地估计单目红外图像的深度信息。

Depth estimation from monocular infrared images based on SVM model

A depth estimation algorithm from monocular infrared image based on nonlinear learning model is proposed in this paper.Firstly,some features corresponding to the depth of infrared image are extracted using stepwise regression and ICA.Secondly,based on the nonlinear SVM model with kernel function,we regress and train the depth features of infrared image using supervised learning method.During the training process,the model parameters are fixed by calculating MMSE from regression analysis of the known data.Finally,we get the depth estimation model which can estimate the depth distribution of monocular infrared image.The experimental results show that most of the depth estimated by the model is consistent with the original depth information of infrared image.