用彩色图像识别田间洋葱和杂草的研究

减少化学产品(化肥、农药等)的用量，是精确农业的主要目标之一。人们力图通过图像分析，测量田间一些地块的杂草和作物的比例并确定杂草的位置，但是在彩色图像里，作物叶子和杂草叶子的交叉给计算机区分它们带来了困难。文中针对两种不同情况，提出了两种不同的识别方法。一种方法是针对没有作物叶子和杂草叶子交叉的情况，另一种方法是针对有某些交叉的情况提出的。前一种方法计算时间短，后一种方法需要大量计算。还提出在图像分割过程中，用不同特征法和两步法进行区域合并的模式，其不同特征包括：区域像素各颜色成分的平均值，区域边界的比值和区域的形状。

Distinguishing Onions and Weeds in Field by Using Color Image

Precision agriculture necessitates the utilization of less quantity of agricultural chemical. By using analysis image, weeds or the proportions of crop and weed leaves can be located in field, but overlapping of crop and weed leaves makes it difficult to distinguish them in color image. In this paper, two different methods are suggested to distinguish onion and weed leaves in two different cases. One method is for non overlapping of onion and weed leaves, the other is for the case with some overlapping. The former method occupies a little of calculating time, the latter method needs a great deal of calculation. A mode of merging regions with two stages and with different attributes is also suggested. The different attributes include: the average of color components of pixels, the rates of frontiers and the shape factor of region.