欢迎访问《图学学报》 分享到：
图学学报

图学学报

• 视觉与图像 • 上一篇    下一篇

基于纹理特征与 HSI 空间的苹果识别与标定

  

  • 出版日期:2016-10-31 发布日期:2016-10-20
  • 基金资助:
    国家自然科学基金项目(61308120)

Apple Identification and Calibration Based on the Texture Characteristics and HSI Space

  • Online:2016-10-31 Published:2016-10-20

PDF (PC)

400

被引次数

6

摘要： 在分析了不同图像分割方法的基础上提出了一种基于颜色特征和纹理特征的图像
分割算法，以解决复杂背景下苹果采摘机器人分割目标与背景的问题。通过分析灰度图像的纹
理特征，求取灰度共生矩阵提取特征，以支持向量机分割图像，并结合HSI 颜色空间的色差特
征达到目标和背景分离的效果。通过与单纯的颜色特征分析和纹理特征分析相比较，该方法在
识别率上高于其他分割算法，同时对于颜色与背景相近的果实也能有很好的分割效果。

关键词: 灰度共生矩阵, HSI 空间, 支持向量机, 分割

Abstract: This paper, after analyzing different image segmentation methods, presents an image
segmentation algorithm based on color and texture features in order to enable an apple picking robot
to separate targets from complex backgrounds. Texture features of grayscale images are analyzed to
obtain ground launched cruise missile for feature extraction and to segment images by support vector
machine. Besides, in combination with chromatic aberration in HSI color space, targets are effectively
separated from their backgrounds. By comparison with the methods of color analysis and texture
analysis, this algorithm provides higher rate of recognition and ever better result in separating fruits
with similar colors to backgrounds.

Key words: gray level co-occurrence matrix, HSI space, support vector machine, segmentation