SEA-LAND SEGMENTATION OF REMOTE SENSING IMAGE BASED ON SPATIAL CONSTRAINT MODEL SUPERPIXEL METHOD

ABSTRACT

None

1.INTRODUCTION

the use of segmention:
1.eliminating the interference of the complex land environment
2.improving near-sea target detection accucy and dfficiency
The superpixel segmentation method is a relatively wide used area segmentation method,which has achieved good results in image segmentation.However, when the background of the coastal zone is more complexm,traditional image segmentation methods,often cannont mask full use of the color feature,texture feature, and spatial information of remote sensing images, and the segmentation effect is not good.

2.RELATED WORK

2.1 SLIC Superpixel Method

Superpixel: It refers to an irregular pixel block with certain visual significance composed of adjacent pixels with similar texture,color,brigthness,and other characteristics.(是指由具有相似纹理、颜色、亮度等特征的相邻像素组成的具有一定视觉意义的不规则像素块)
It uses the similarity of feature between pixels to group pixels.It replaces a large number of pixels with a small number of superpixwls to express image features,which significantly reduces the complexity of image post-precessing, so it is usually used as a preprocessing step of the segmentation algorithm.
According to the classification of implementation methods.mainstream superpixel mehods are divided into graph theory and gradient descent.
The implementation steps of the SLIC algorithm are as follows:
(1)假定图片有总共N个像素点，将这N个像素点被分割为K个superpixel，所以对于每个superpixel的大小是N/K，两个临近的种子点大约是S=sqrt(N/K);
(2)计算$3\times 3$种子点周围的梯度值，同时移动种子点移到领域中梯度最小的位置，避免落在梯度较大的边界上，影响聚类效果的问题。
（3）在每一个$2S\times 2S$领域包围了种子点，同时将category labels对应到pixels；
（4）将空间和亮度距离进行计算。计算如下图所示：

(5)重复上文的步骤直到收敛。

3.PROPOSED METHOD

本文提出的方法有四个主要的步骤：
（1）计算Search point和seed point之间的距离计算直接使用平方来计算，不不使用开根号？？？？（这tm也算创新）计算公式如下：

（2）提取特征和颜色特征
（3）基于RAG理论，每个区域获得超像素分割
（4）计算机构生成