变化光照下树上柑橘目标检测与遮挡轮廓恢复技术

遮挡是自然场景中普遍存在的问题,为在变化光照条件下正确检测出自然环境中的树上成熟水果目标,从而为全天候的机械采摘提供运动参数,研究了基于彩色信息和目标轮廓整合的树上遮挡柑橘检测方法。在对自然光照条件下的可见光彩色图像进行颜色特征分析的基础上,建立了利用R-B色差图融合归一化RGB颜色空间的方法,对树上水果目标区域进行了初分割。然后提取R-B色差图的主边缘构造边缘片段集,根据边缘片段长度、弯曲程度以及凹凸性来选择有效边缘片段集,对每个有效边缘片段进行拟合,最后根据水果形状知识选择出有效拟合目标椭圆。根据对不同光照条件和遮挡程度的场景处理的结果表明,所提算法能有效恢复出树上存在遮挡的水果目标,最后遮挡轮廓恢复结果的相对误差为5.34%。

Detection of Citrus Fruits within Tree Canopy and Recovery of Occlusion Contour in Variable Illumination

A method based on color information and contour fragments was developed to identify citrus fruits in variable illumination conditions in the tree canopy, in order to guide the robots for harvesting citrus fruits. The color properties of target objects within natural citrus-grove scenes under various light conditions were analyzed, and a preliminary segmentation was put forward by fusing the Chromatic aberration information and normalized RGB model. The set of contour fragments was constructed via detecting the significant edge of Chromatic aberration map of R and B channels. The valid subset was selected by three parameters of the frament: length, bending degree and concavo-convex geometry characteristic. The ellipse fitting procedure was done to every frament, and the valid ones were chosen by the knowledge of fruit shape. The results showed that the occlusion contour were effectively recoveried under various light conditions using the proposed method, and the relative error of occlusion recovery was 5.34%.