基于色差原理与双目测距技术的红花识别与定位方法

Safflower Identification and Positioning Method based on Chromatic Aberration Principle and Binocular Ranging Technology

为了在图像中识别和提取红花并获得物理空间下的位置信息,提出一种基于色差原理与双目测距技术的红花识别与定位方法。首先,通过双目深度相机采集红花图像;然后利用改进的色差分割法将红花从背景中分割出来,求出红花区域轮廓二维重心点坐标;最后利用双目深度相机,根据二维轮廓重心点坐标,获取该坐标点下红花距离相机深度以及三维坐标信息,再通过坐标系之间的转换将其传递给机械臂,完成采摘工作。试验表明,在纵向Z方向,深度值在300~900 mm范围内,最大绝对误差为8.62 mm,最大相对误差为0.95%;在左右X方向,最大绝对误差为3.95 mm,最大相对误差为4.05%;在上下Y方向,最大绝对误差为2.95 mm,最大相对误差为3.37%,测量精度可靠准确,满足采摘需求。

In order to identify and extract the safflower in the image and obtain the position information in the physical space,a safflower recognition and location method based on the chromatic aberration principle and binocular ranging technology was proposed.Firstly,the safflower image was collected by binocular depth camera;then the improved color difference segmentation method was used to segment the safflower from the background,and the two-dimensional center of gravity coordinates of the safflower contour were obtained;finally,the binocular depth camera was used to obtain the safflower distance from the camera depth and three-dimensional coordinate information under the coordinate point according to the coordinates of the two-dimensional contour gravity center point,and then it was transmitted to the manipulator through the conversion between coordinate systems to complete the picking work.The test shows that in the longitudinal Z direction,the maximum absolute error and relative error are 8.62 mm and 0.95%respectively when the depth value is in the range of 300~900 mm;in the left and right X direction,the maximum absolute error is 3.95 mm and the maximum relative error is 4.05%;In the up-down Y direction,the maximum absolute error is 2.95 mm and the maximum relative error is 3.37%.The measurement accuracy is reliable and accurate to meet the picking requirements.