基于图像处理的甘蔗茎节识别与蔗芽检测

Sugarcane Stem Section Identification and Sugarcane Bud Detection Based On Image Processing

为实现机器能够准确无误的找到甘蔗蔗芽的位置,让切刀在指定的位置切割甘蔗,运用图像处理技术对甘蔗的茎节和蔗芽分别进行了定位识别。通过滤波去噪、颜色空间转换、图像融合获得图像的二值化信息;再通过LBP(Local Binary Pattern)纹理特性和形态学操作实现甘蔗区域的分割。之后,通过计算每个茎节上列像素值之和,统计出位于直线中线的像素值,即该像素值对应的列坐标为茎上的最优直线。最后,由蔗芽总是环绕茎节分布的特征,结合bounding Rect函数,定位出蔗芽的位置。试验结果表明:该算法的甘蔗茎节识别率,在单节时能达到100%,在多节时稳定在80%以上,算法执行时间为0.507s。

In order to locate the cane buds accurately and correctly and enable the cutter to work at the designed position.The image processing technology was used to locate and identify the stem segments and sugarcane buds of sugarcane.The binary information of image was obtained through filter denoising,color space conversion and image fusion;then the segmentation of sugarcane area is realized by LBP(Local Binary Pattern)texture characteristics and morphological operations.Then,by calculating the sum of the column pixel values on each stem segment,the pixel value of the line in the straight line is counted,that is,the column coordinate corresponding to the pixel value is the optimal straight line on the stem.Finally,the location of cane bud was determined by the characteristics of the cane bud distribution around the stem node and boundingRect function.The experimental results show that the recognition rate of sugarcane stem section can reach up to 100%in single section and more than 80%in multi-section,and the algorithm execution time is 0.507 s.