基于OpenCV的自适应阈值图像前景提取  

Prospect Extraction of Adaptive Threshold Image Based on OpenCV

作　　者：胡宗进[1] 范文强[1] 于光玉 刘宝军[1] 

机构地区：[1]烟台大学,山东烟台264000 [2]聊城大学,山东聊城252000

出　　处：《光电技术应用》2017年第1期56-59,共4页Electro-Optic Technology Application

摘　　要：植株的图像提取是进行植株位置判定的基础,是图像分割的范畴。图像的色彩信息是不容忽视的一方面,基于此提出了一种基于色彩的区域分割方法。由于植株的绿色成分较大,故可采用最大分量提取的方法实现对图片中植株部分的分割,此种方法计算量很小,只需按像素操作即可。另外其中提出的非彩色部分剔除方法能够较好的分离出建筑物和植物,使得分割出的图像更为准确。自适应阈值二值化计算方法可以根据图像自身的亮度值进行实时计算,能够较好的适应不同光照条件下图片的二值化处理。经过实例验证,此方法具有一定的有效性和合理性。Plant image extraction is the basis for determining the plant location, which is in the field of image segmentation. The image color information cannot be ignored on one hand, based on this point of view, a colored region segmentation method is proposed. Because of the large green component of the plant, it is possible to use the maximum component extraction method to realize the segmentation of the plant in the image. The method has minimum calculation and operates only on pixels. In addition, the method can not only separate the buildings and plants, but also make the image more accurate. Adaptive threshold binaryzation calculation method can be used at real time according to the brightness of the image and can be better adapted to image binaryzation processing at dif- ferent light condition. The examples prove that the method is effective and reasonable.

关 键 词：彩色图像分割 植株提取 图像二值化 最大分量提取 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]