激光传感器在喷雾靶标检测中的研究应用

发布时间：2018-03-17 12:45

本文选题：现代化农业　切入点：变量精密施药　出处：《农业现代化研究》2017年02期 　论文类型：期刊论文

【摘要】：树木尺寸、形状等特征信息的获取,是现代化农业中重要的环节,为农林管理中的变量精密喷雾提供依据。为了实现对树木靶标信息的精确测量,搭建了基于UTM-30LX型激光传感器的室内滑台靶标检测系统,实现对规则物体、仿真树的扫描检测,通过编写的C++程序,实时读取数据信息并存储于控制计算机中。利用MATLAB软件对存储的数据信息进行后期的离线分析,重构物体三维图像,通过最大相对误差、轮廓相似度两个参数,分析激光传感器检测得到靶标目标尺寸、重构目标三维图像的精确度。以长方体柜子、圆柱体泡沫、2棵仿真树为试验对象,按照选取的3个检测距离和4个检测速度分别对4个试验对象进行检测与数据分析。试验结果显示,多次试验下,与人工测量尺寸比较而言,激光传感器测得的长方体柜子长度、宽度和高度尺寸最大相对误差分别为6.76%、6.86%和3.92%,圆柱体泡沫的高度、直径尺寸最大相对误差分别为4.25%和7.33%,仿真树1高度、宽度尺寸最大相对误差分别为4.06%和5.91%,仿真树2高度、宽度尺寸最大相对误差分别为3.24%和4.00%;仿真树1、2的激光扫描重构图像与对应实物轮廓相似度最小值分别为0.928 8和0.932 6。研究表明,UTM-30LX型激光传感器在各种试验条件下对多种试验靶标特征检测均具有较好的精确性和准确性,可以担负变量喷雾检测的任务,具有应用到农业生产中的前景。
[Abstract]:The acquisition of characteristic information, such as tree size and shape, is an important link in modern agriculture, which provides the basis for variable precision spray in agroforestry management. Based on UTM-30LX laser sensor, an indoor target detection system for sliding platform is built, which can scan and detect regular objects and simulation trees. The data information is read in real time and stored in the control computer. The MATLAB software is used to analyze the stored data information off-line, and to reconstruct the 3D image of the object. The maximum relative error and the contour similarity are two parameters, the maximum relative error and the contour similarity. The target size is obtained by laser sensor detection, and the accuracy of 3D image reconstruction is obtained. Two simulation trees, cuboid cabinet and cylindrical foam tree, are used as experimental objects. According to the selected three detection distances and four detection velocities, the four test objects were tested and the data were analyzed respectively. The test results show that, after many tests, the size of the four tested objects is compared with the manual measurement. The maximum relative error of cuboid cabinet length, width and height measured by laser sensor is 6.76% and 3.92%, respectively. The maximum relative error of cylindrical foam height and diameter dimension is 4.25% and 7.33, respectively. The maximum relative error of width dimension is 4.06% and 5.91 respectively. The maximum relative error of width dimension is 3.24% and 4.00, respectively, and the minimum similarity between the reconstructed image and the corresponding contours is 0.928 8 and 0.932 6, respectively. The results show that the laser sensor of UTM-30LX is suitable for different experimental conditions. Many kinds of test target feature detection have good accuracy and accuracy. It can take on the task of variable spray detection, and has the prospect of application in agricultural production.
【作者单位】：江苏大学电气信息工程学院;
【基金】：国家自然科学基金项目(51505195) 江苏省自然科学基金项目(BK20130501) 江苏省高校自然科学基金(13KJB210002)~~
【分类号】：TP391.41;TP212

【参考文献】