大田环境下智能移动喷药机器人系统研究
发布时间：2018-12-12 22:35
【摘要】：随着农业机械化和机器人智能化的发展,现代农业机械发展方向已经逐步转向智能农业、设施农业和高效农业。高效、智能和精细成了现代农业机械的发展目标。这一学科集合了机械、电子、控制理论以及植物保护等诸多领域最前沿的研究成果。一定程度上小型移动热雾机在病虫害防治方面填补了玉米、高粱、甘蔗等高大作物中后期难以防治的空白,具有极大的发展空间和研究意义。大田环境下病虫害的防治是现代农业研究的核心问题之一。为了实现小型移动热雾机器人在田间自动实现农药的均匀喷洒,本文以机械化种植并生长到中后期的玉米为研究对象,针对移动机器人在作物行间大面积均匀施药展开核心技术研究,自主研发了作物行间小型移动喷雾机器人,对机器人移动中喷雾的雾滴沉积均匀性进行了研究。针对多行均匀喷雾过程中机器人速度保持下的自主行走的问题,提出了基于径向基函数的单目视觉路径规划方法,分析了在不同移动速度下多行喷雾的雾滴沉积特性,并进行了试验和分析。本文主要研究内容和结果如下：(1)针对热雾机工作原理进行分析,分解热雾机喷雾过程。通过将热雾机喷雾到雾滴沉积在作物表面的过程分解为药液的雾化和雾滴在空气中运动至沉积位置两步骤,根据前人的研究成果,分析热雾机实现大面积喷雾的机理。选用同型号不同孔径喷药嘴时,雾滴最终沉积趋势相似,然而在施药有效距离、平均粒径和覆盖率上仍有差异。(2)针对搭载于移动平台的喷雾系统在不同移动速度下的雾滴分布均匀性、覆盖率和有效距离这三个热雾机工作效率的重要指标。本文提出了一种在大田实验中水敏试纸半定量研究雾滴沉积分布的方法。水敏试纸中反应区域通过450-455nm波长蓝色圆周阵列的光源的照射与试纸本底进行区别,减小了杂散光的干扰,提高了所采集图像二值化分割的效果。引入置信区间估算的方法去除采集数据中的噪声数据。对图像处理的数据可信区间上下限分析进一步去减低系统误差,提高检测精度。此方法对于雾滴的覆盖率计算的系统重复测量误差低于5%；当雾滴覆盖率在10%-20%时,对于雾滴Dv0.5系统重复测量误差也同样低于5%。利用此方法对机器人喷雾系统在不同速度下雾滴沉积的特性进行研究,分别进行雾滴分布均匀性、覆盖率和有效距离的分析。测绘了热雾机移动过程中雾滴覆盖率的分布曲线。(3)研究了机器人在狭小作物行间基于单目视觉的导航方法。通过采取增强光照的方法来削弱光照不均对结果的影响。通过对秸秆图像块的主成分分析、形态学分割和相机的标定,实现了图像中秸秆根部位置的定位;最后利用简化的径向基函数算法根据所得秸秆位置坐标规划行走路径,提高了秸秆块提取的容错能力。采用本系统的导航方法,机器人保持速度为0.6m/s时,与秸秆的最大碰撞率为6.38%。系统的图像处理平均耗时220ms,路径规划的平均耗时为30ms,可以满足运行速度1.2m/s以下的直线自主行驶的响应需求。最后,本文实验样机在涡阳及怀远实验田中进行了多行作业的雾滴沉积研究,测试了在不同速度下雾滴沉积的均匀性和覆盖率,实现了精准施药的优化,实验得到样机的喷雾效率曲线并对各因素进行针对性的分析。研究结果为进一步实现小型移动热雾机器人在大田精准施药的方法的研究提供了实验依据,为提高农药的喷洒效率提供了实验参考。
[Abstract]:With the development of agricultural mechanization and robot, the development direction of modern agricultural machinery has gradually turned to intelligent agriculture, facility agriculture and high-efficiency agriculture. High-efficiency, intelligent and fine is the development goal of modern agricultural machinery. The subject has the most advanced research results in the fields of machinery, electronics, control theory and plant protection. The small-sized mobile hot-fog machine has the great development space and research significance in the aspect of preventing and controlling the diseases and insect pests of the large crops such as corn, sorghum, sugar cane and the like. The prevention and control of diseases and insect pests in the field of field is one of the core problems of modern agricultural research. in ord to realize that uniform spraying of the small-scale mobile hot-fog robot in the field, In this paper, the small-scale moving spray robot between the interline of the crop is researched and developed, and the uniformity of the droplet deposition in the motion of the robot is studied. In order to solve the problem of autonomous walking in the process of multi-row uniform spraying, a single-order vision path planning method based on radial basis function is proposed, and the droplet deposition characteristics of multi-row spray at different moving speeds are analyzed, and the test and analysis are carried out. The main contents and results of this paper are as follows: (1) The working principle of the hot fog machine is analyzed, and the spraying process of the hot fog machine is decomposed. The mechanism of the large-area spraying of the hot-fog machine is analyzed by the process of spraying the hot-fog machine to the surface of the crop and decomposing into the atomization of the liquid medicine and the movement of the fog-fog droplets in the air to the deposition position. The final deposition tendency of the fog drops is similar when the same model of different aperture spray nozzle is selected, however, the effective distance, the average particle size and the coverage rate of the pesticide application are still different. and (2) the distribution uniformity, coverage and effective distance of the fog drops mounted on the moving platform at different moving speeds are important indexes of the working efficiency of the three hot-fog machines. In this paper, a method for semi-quantitative study of the droplet deposition profile of water-sensitive test paper in field experiment is presented. The reaction area in the water-sensitive test paper is different from the background of the test paper through the illumination of the light source of the blue circumferential array with the wavelength of 450-455nm, so that the interference of the stray light is reduced, and the effect of the two-value division of the acquired image is improved. The method of introducing the confidence interval estimate removes the noise data from the acquisition data. the lower limit analysis on the data credible interval of the image processing further reduces the system error and improves the detection accuracy. The measurement error of the system is less than 5% for the coverage of the fog drops; when the coverage of the fog drops is between 10% and 20%, the measurement error of the system is also less than 5% for the drop Dv0.5 system. The characteristics of the droplet deposition at different speeds were studied by this method, and the distribution uniformity, coverage and effective distance were analyzed respectively. and the distribution curve of the coverage of the fog drops during the movement of the hot fog machine is mapped. (3) The navigation method based on monocular vision in the interline of small crop is studied. the effect of uneven illumination on the results is reduced by adopting a method of enhancing the illumination. By analyzing the principal component of the straw image block, the morphological segmentation and the calibration of the camera, the positioning of the position of the straw root in the image is realized; and finally, a simplified radial basis function algorithm is utilized to plan the walking path according to the position coordinate of the obtained straw, and the fault-tolerance capability of the straw block extraction is improved. With the navigation method of the system, the maximum collision rate of the straw is 6.38% when the speed of the robot is 0. 6m/ s. The average time of the image processing of the system is 220ms, the average time of the path planning is 30ms, and the response requirement of the straight-line autonomous driving with the running speed of 1. 2m/ s can be met. In the end, the experimental prototype of this paper has carried on the droplet deposition of multi-line operation in the eddy-Yang and Huaiyuan experimental field, tested the uniformity and coverage of the droplet deposition at different speeds, and realized the optimization of the precise application. The spray efficiency curve of the prototype was obtained and the factors were analyzed. The results of the study provide an experimental basis for the further research on the method of precision application of the small moving hot-fog robot in the field, and provides an experimental reference for improving the spraying efficiency of the pesticide.
【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP242
，
本文编号：2375354
[Abstract]:With the development of agricultural mechanization and robot, the development direction of modern agricultural machinery has gradually turned to intelligent agriculture, facility agriculture and high-efficiency agriculture. High-efficiency, intelligent and fine is the development goal of modern agricultural machinery. The subject has the most advanced research results in the fields of machinery, electronics, control theory and plant protection. The small-sized mobile hot-fog machine has the great development space and research significance in the aspect of preventing and controlling the diseases and insect pests of the large crops such as corn, sorghum, sugar cane and the like. The prevention and control of diseases and insect pests in the field of field is one of the core problems of modern agricultural research. in ord to realize that uniform spraying of the small-scale mobile hot-fog robot in the field, In this paper, the small-scale moving spray robot between the interline of the crop is researched and developed, and the uniformity of the droplet deposition in the motion of the robot is studied. In order to solve the problem of autonomous walking in the process of multi-row uniform spraying, a single-order vision path planning method based on radial basis function is proposed, and the droplet deposition characteristics of multi-row spray at different moving speeds are analyzed, and the test and analysis are carried out. The main contents and results of this paper are as follows: (1) The working principle of the hot fog machine is analyzed, and the spraying process of the hot fog machine is decomposed. The mechanism of the large-area spraying of the hot-fog machine is analyzed by the process of spraying the hot-fog machine to the surface of the crop and decomposing into the atomization of the liquid medicine and the movement of the fog-fog droplets in the air to the deposition position. The final deposition tendency of the fog drops is similar when the same model of different aperture spray nozzle is selected, however, the effective distance, the average particle size and the coverage rate of the pesticide application are still different. and (2) the distribution uniformity, coverage and effective distance of the fog drops mounted on the moving platform at different moving speeds are important indexes of the working efficiency of the three hot-fog machines. In this paper, a method for semi-quantitative study of the droplet deposition profile of water-sensitive test paper in field experiment is presented. The reaction area in the water-sensitive test paper is different from the background of the test paper through the illumination of the light source of the blue circumferential array with the wavelength of 450-455nm, so that the interference of the stray light is reduced, and the effect of the two-value division of the acquired image is improved. The method of introducing the confidence interval estimate removes the noise data from the acquisition data. the lower limit analysis on the data credible interval of the image processing further reduces the system error and improves the detection accuracy. The measurement error of the system is less than 5% for the coverage of the fog drops; when the coverage of the fog drops is between 10% and 20%, the measurement error of the system is also less than 5% for the drop Dv0.5 system. The characteristics of the droplet deposition at different speeds were studied by this method, and the distribution uniformity, coverage and effective distance were analyzed respectively. and the distribution curve of the coverage of the fog drops during the movement of the hot fog machine is mapped. (3) The navigation method based on monocular vision in the interline of small crop is studied. the effect of uneven illumination on the results is reduced by adopting a method of enhancing the illumination. By analyzing the principal component of the straw image block, the morphological segmentation and the calibration of the camera, the positioning of the position of the straw root in the image is realized; and finally, a simplified radial basis function algorithm is utilized to plan the walking path according to the position coordinate of the obtained straw, and the fault-tolerance capability of the straw block extraction is improved. With the navigation method of the system, the maximum collision rate of the straw is 6.38% when the speed of the robot is 0. 6m/ s. The average time of the image processing of the system is 220ms, the average time of the path planning is 30ms, and the response requirement of the straight-line autonomous driving with the running speed of 1. 2m/ s can be met. In the end, the experimental prototype of this paper has carried on the droplet deposition of multi-line operation in the eddy-Yang and Huaiyuan experimental field, tested the uniformity and coverage of the droplet deposition at different speeds, and realized the optimization of the precise application. The spray efficiency curve of the prototype was obtained and the factors were analyzed. The results of the study provide an experimental basis for the further research on the method of precision application of the small moving hot-fog robot in the field, and provides an experimental reference for improving the spraying efficiency of the pesticide.
【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP242
，
本文编号：2375354
本文链接：https://www.wllwen.com/shoufeilunwen/xxkjbs/2375354.html
