基于光度立体视觉的蔬菜秧苗叶片形态测量方法  

Measurement Method of Vegetable Seedling Leaf Morphology Based on Photometric Stereo

作　　者：冯青春[1,2] 陈建[2] 李翠玲[1,3] 范鹏飞 王秀[1,4] FENG Qingchun1,2 CHEN Jian2 LI Cuiling1,3 FAN Pengfei1,4 WANG Xiu1,4(1. Beijing Research Center of Intelligent Equipment for Agriculture, Beijing 100097, China;2. College of Engineering, China Agricultural University, Beijing 100083, China ;3. National Research Center of Intelligent Equipment for Agriculture, Beijing 100097, China ;4. Beijing Key Laboratory of lntelligent Equipment Technology for Agriculture, Beijing 100097, Chin)

机构地区：[1]北京农业智能装备技术研究中心,北京100097 [2]中国农业大学工学院,北京100083 [3]国家农业智能装备工程技术研究中心,北京100097 [4]农业智能装备技术北京市重点实验室,北京100097

出　　处：《农业机械学报》2018年第5期43-50,共8页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金项目(61703048);北京市优秀人才培养资助青年骨干个人项目(2015000020060G134);北京市农林科学院青年科研基金项目(QNJJ201722)

摘　　要：为了实现蔬菜秧苗长势智能化在线评估,设计了基于光度立体视觉的蔬菜秧苗叶片曲面形态测量方法和装置,用于对叶片倾角、长度和面积进行精确测量。以15~30 d苗龄辣椒秧苗为测量对象,构建了针对其冠层叶片形态测量的试验装置。根据4组光源在标定球表面成像的反射关系组合方程,标定其对秧苗叶片的照射向量;以D65白色标准板为参考,采用图像RGB分量线性矫正的方法,对不同方位光源辐射强度差异进行补偿,克服光源强度波动对叶片图像明暗信息的影响。根据叶片在不同方位光源照射下图像的明暗特征,基于光度立体视觉获取主叶脉区域离散梯度信息,在此基础上采用最小二乘方法拟合叶片空间平面,以恢复叶片空间倾斜信息,进一步结合叶片图像尺寸测算其实际长度和面积。试验结果表明,视觉系统对秧苗倾角、叶片长度和面积的测量结果与人工测量结果相比平均偏差分别为6.29°、3.82 mm、56.53 mm2,叶片长度和面积与人工测量结果的决定系数R2分别为0.936 3、0.866 4,且对于叶片伸展充分、无遮挡的幼龄秧苗测量精度较高,可为进一步开发温室苗床秧苗长势在线监测设备提供重要技术支撑。In order to achieve the on-line nondestructive assessment for vegetable seedling growth,the method and device for measuring the seedling ＇s leaf spatial morphology based on photometric stereo vision was introduced,and the morphology information,including seedling leaf ＇s inclination angle,length and size. The 15 ~ 30 d pepper seedlings were adopted as research objects,and the test device for measuring their canopy leaves was built,which involved the four different LED lights ＇ orientation calibration and the image color＇s linear correction referring to the D65 palette＇s color variation under the various illumination. According to the light reflexive relation on the calibration ball from the LED lights to the camera,the spatial orientations of illuminants under the camera coordinate were determined. And the image color correction was supposed to overcome the color distortion caused from the lights＇ luminance fluctuation. According to the seedling leaves＇ light-shadow feature under various illuminants,the discrete gradient information along the leaf＇s main vein was obtained based on photometric stereo vision. And referring to the gradients vectors,the leaf＇s space plane was fit by adopting the least-square principle.The leaf＇s inclination angle was measured as the angle between the fitting plane and camera imaging plane,thus the actual spatial leaf size also could be accurately measured on the basis of the image size.As the test result shown,compared with the manual measurement result,the auto-measurement method had an average errors of 6. 29° for leaf inclination angle,3. 82 mm for leaf length,and 56. 53 mm2 for leaf size,and the determination coefficients on the length and leaf size measurement were 0. 936 3 and0. 866 4,respectively. Besides,the measuring result would be more accurate for the younger seedling with fully-opened and unoverlapped leaves.

关 键 词：蔬菜秧苗 长势评估 光度立体 叶片倾角 叶片面积 平面拟合 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论] S126[自动化与计算机技术—计算机科学与技术]