基于Hopfield神经网络的FLIR图像分割
Segmentation of FLIR Images by Hopfield Neural Network
-
摘要: 针对前视红外(FLIR)图像的分割,在基于模型的FLIR图像分割算法所提出的全 局准则函数及初始概率确定方法的基础上.建立了与之相对应的Hopfield网络的能量函数 及网络的初始状态,当网络运行达到稳定状态后,便可获得图像的分割结果.分析了能量函数 中,目标函数与约束条件的加权系数对分割结果的影响,并根据分割结果的非模糊性准则,提 出了一个确定加权系数的、简单有效的方法.给出了针对真实红外目标图像的分割结果.
-
关键词:
- 图像分割 /
- 梯度松弛技术 /
- 前视红外图像 /
- Hopfield模型
Abstract: On the basis of the global criterion and the initial assignment of probabilities in model-based forward-looking infrared (FLIR) images segmentation algorithm, we establish the corresponding energy function and initial state of Hopfield neural network (HNN). When HNN converges, the segmentation result can be obtained. The effect of the weights of the objective function and the constraint condition in the energy of HNN is analyzed, and based on the nonamblguity criterion, a simple and effective method is proposed to determine the apropriate weights. Experimental results with real FLIR images are given.-
Key words:
- Image segmentation /
- gradient relaxation technique /
- FLIR image /
- Hopfield model
-
计量
- 文章访问数: 2530
- HTML全文浏览量: 277
- PDF下载量: 1053
- 被引次数: 0
下载:
