DEVELOPMENT OF AUTOMATIC ROW-FOLLOWING FIELD MACHINERY

为了研制一种自动跟行的田间作业机械，研制了作物行探测器和自动转向控制方法。研究结果如下：1)建立了一种新的检测方法。与普通图像处理方法相比，该方法可以更快地检测出作物行的偏移量和行距。2)试制了三种作物行探测器。现场测试的结果表明，每台探测器的精度都达到了实际使用要求。特别是采用二维彩色图像传感器的探测器，由于对天空和背景颜色条件的适应性，以及探测算法的高度灵活性，是最实用的。该探测器的平均检测误差为()Sy。-.3)行跟随控制实验结果表明，缓冲转向和有意识的纠偏控制对弯行是有效的。4)为了适应不同类型的作物行，需要改进检测方法以能够检测到弯曲行，以及作物行的海峡部分和弯曲部分的交替控制技术。

In order to develop an automatic row-following field machinery, Crop-row detector and the methods of automatic steering control have been developed. The results obtained are as follows ;1) The new detecting method was developed. Offset and heading to the crop-row can be detected faster by this method than by ordinary image processing.2) Three types of Crop-row detector were made on trial basis. As results of the field tests, it was concluded that each detector was accurate enough for practical use. Especially, the detector with 2D color image sensor was most practical, because of adaptability to sky and background color conditions, and high flexibility of detecting algorithm. The average detecting error of this detector were (]SY.+-。[)2cm in offset and 0.5 degree in heading detection.3) As a result of experiment on row following control, the buffered steering and the control of offset correction conscious was effective to the curved row. The defects of delayd correction by buffered steering in a strait row and the frequent steering, however, should be improved.4) Improvements in detecting method to be able to detect a curved row and alternation control technique in a strait part and a curved part of crop row would be necessary for adaptation to a various type of crop row.