智能高效株间锄草机器人研究进展与分析

非化学方式除草是摒弃除草剂、生产有机农产品的前提,传统的中耕锄草机主要解决行间锄草,由于株间苗草集聚,机械锄草难度较大,目前主要依靠人工,劳动成本高且效率低。智能株间锄草机器人是一种能够实时识别作物行和苗草信息,并能控制株间锄草刀高速作业的自动锄草装备,具有智能、高效、环保等特点,可大大减少劳动力,提高锄草效率,降低锄草成本。该文主要对近年来国外研究较为成熟的株间锄草机器人进行介绍,概述了中国该方面的研究进展,对苗草信息获取、对行、锄草装置、驱动方式、时速等几个技术点进行分析和比对,提出了如何提高信息获取速度,增强系统实时性,以及如何改进机器视觉标定方法,提高苗草定位准确性是苗草信息获取技术存在的关键问题,强调了锄草装置在系统中的重要性;针对不同形态作物、不同土质土壤研制针对性强、锄草效果好的锄草装置是锄草机器人的基础,同时由于系统集成性及动力系统与速度匹配仍无法满足田间高负载、高速的锄草作业要求,因此加强该方面研究力度,研制使用性强、效率高的株间锄草机器人仍是中国的研究重点和方向。最后,提出多传感器融合、模块化、小型化的株间锄草机器人将是未来发展趋势,是实现中国农业有机、精准、高效生产的重要依据。

Study review and analysis of high performance intra-row weeding robot

Abstract: To ensure food security and sustainable development of agriculture, it is critically important to develop organic agriculture and pollution-free agricultural products. Non-chemical weed control is important for organic agricultural production. Intra-row weeds are more difficult to be eliminated than inter-row weeds due to their proximity to the crop or seed line. Therefore, techniques for intra-row weed control remain to be one of the biggest challenges today. Efficient weeding robots can differentiate and remove inter-and intra-row weeds simultaneously, while identify and provide species information. The researches of intra-row weed control are more common in Europe because of their political interest and a push from the market. Nowadays, other countries, like the United States, Japan, Canada and China had carried out the researches in this area. This review describes the current state of the art of efficient weeding robots in detail to demonstrate the potential of the technology in the field. Several key technologies of weed control systems have been analyzed and compared. Three core technologies were summarized: guidance, detection and identification, and intra-row device. Row guidance which can use machine vision for crop row detection or GPS is an essential technology for weeding robot. Detection and identification which can identify and locate the crops in real time by three main approaches (computer vision, GPS and proximity sensor) is a critical technology. Furthermore, fusion of computer vision and GPS is seen today as the most promising strategy because advantages and disadvantages of absolute and relative referencing principles compensate each other. Intra-row device divided into rotation-type and swing type according to movements is a foundation of weed control. China is much later than the other developed countries in this area, but it has a booming development and grows rapidly. However, several technical problems exist as follows: 1) Recognition algorithm based on machine vision takes longer time, which makes the real-time of the system perform poorly; 2) Visual calibration method is cumbersome and complex under the non-structural condition; 3) The importance of intra-row weeding device development is ignored, weeding actuators directly affect weeding effect, the number and flow conditions of soil surrounding crops; 4) The hydraulic oil source is contaminated easily because tractor hydraulic oil is mainly used for lubrication; 5) Electric and pneumatic weeding robots are limited by poor battery life, and thus it is not available for field work; 6) The integrated performance of weed control system cannot meet the requirement of heavy load and efficient weeding in the field, and there still are some gaps between prototype and actual application; 7)There are less study on matching power system with speed of weeding robot that is the key to improve weeding precision and speed. According to the review, the further researches on weeding robots focusing on sensor fusion, efficient work, modular, miniaturization, high stability, and long life time have been proposed. With China gradually losing the advantage of cheap labor as well as sustained reduction of agricultural personnel, intra-row weeding robots are going to play an important role in agricultural production in China.