Lab色彩模型下梯度Hough圆变换的成熟草莓识别

自然生长状态下的草莓,果实密集,形状多变,枝叶遮挡率较高,使用机器采摘对成熟草莓的图像识别具有一定的难度。针对这个问题,提出了Lab色彩模型下采用梯度Hough圆变换的成熟草莓识别方法。以自然生长状态下的草莓图像为研究对象,在Lab色彩模型空间里对图像分割及阈值化并采用梯度Hough圆变换等方法,识别成熟草莓并计算出草莓中心。试验结果表明,对随机选取的100张草莓图片,该方法识别成熟草莓的相对偏差为1.07%,对枝叶遮挡、成熟草莓相互遮挡等情况,具有较好的识别效果。与传统的Hough变换相比,这一方法节省了运行内存,提高了时间效率,能够满足机器采摘对成熟草莓识别率与定位的要求。

Ripe strawberry recognition based on gradient hough circle transform in lab color model

Strawberries in natural growth state have characteristics of dense fruits, different shapes and high occlusion rate of branches and leaves. It is difficult to recognize ripe strawberries when implementing machine picking. In order to solve this problem, this paper proposes one approach for identifying ripe strawberries among the strawberry image in natural growth state using gradient Hough circle transformation under Lab color model. In Lab color model space, image segmentation and thresholding are performed, and gradient Hough circle transformation is used to identify ripe strawberries and calculate strawberry center. Experimental results showed that the relative deviation of identifying ripe strawberries is 1.07% for 100 randomly selected strawberry images, and it has good ability of dealing with recognition effect for branch and leaf occlusion, mutual occlusion of ripe strawberries. Compared with the traditional Hough transform, the proposed method has less demand of running memory and time, consequently, it can meet the recognition and location requirements of machine picking for ripe strawberries.