基于HSV空间再生稻植株与土壤背景图像分割

发布时间：2018-07-03 04:20

本文选题：再生稻 + 农田环境　；参考：《农机化研究》2017年07期

【摘要】：针对再生稻收割机视觉导航的稻田图像分割问题,结合再生稻植株的生长特点和再生稻避儕的要求,利用相机于农田采集再生稻图片,结合RGB、HSV、YCr Cb空间中的常用灰度化因子,进行灰度化对比试验并分析其直方图特征,得出在HSV空间的S分量灰度化;采用最大类间方差法(Otsu)得到初步分割阈值T,经进一步分析为保留较完整的不同成熟度再生稻植株特征,加入修正因子-a得到阈值T-a对图像二值化;再通过数学形态学,面积法过滤等后续处理,形成收割机行走的左右边界区域。结果表明:处理1副像素419×310的图像平均耗时0.053 s,可满足今后的实时性要求,分割出的图像基本上反应了再生稻植株的走势特征,与人眼判断植株边缘位置基本相符合。
[Abstract]:Aiming at the problem of rice field image segmentation based on visual navigation of ratooning rice harvester, combined with the growth characteristics of ratooning rice plant and the requirements of rooting rice peer-avoidance, the paper used camera to collect the images of ratooning rice in farmland, and combined with the commonly used grayscale factors in the space of RGB HSV and YCr CB. The contrast test of grayscale and the analysis of histogram feature are carried out, and the S component grayscale in HSV space is obtained. The initial segmentation threshold T was obtained by using the maximum inter-class variance method (Otsu). After further analysis, the threshold value T-a was obtained by adding the correction factor -a to preserve the complete plant characteristics of different maturity ratooning rice, and then the binary value of the image was obtained by mathematical morphology. Area filtering and other follow-up processing to form the left and right border area of the harvester walking. The results show that the average processing time of 1 pixel 419 脳 310 image is 0.053 s, which can meet the real-time requirements in the future. The segmented image basically reflects the trend characteristics of the regenerated rice plant and is basically consistent with the human eye in judging the edge position of the plant.
【作者单位】：福建农林大学机电工程学院;
【基金】：福建省自然科学基金项目(2016J01701) 福建农林大学机械工程学科整体学科水平提升计划项目(612014049)
【分类号】：TP391.41

【相似文献】