采用双模态联合表征学习方法识别作物病害

基于深度卷积神经网络的视觉识别方法在病害诊断中表现出色,逐渐成为了研究热点。但是,基于深度卷积神经网络建立的视觉识别模型通常只利用了图像模态的数据,导致模型的识别准确率和鲁棒性,都依赖训练数据集的规模和标注的质量。构建开放环境下大规模的病害数据集并完成高质量的标注,通常需要付出巨大的经济和技术代价,限制了基于深度卷积神经网络的视觉识别方法在实际应用中的推广。该研究提出了一种基于图像与文本双模态联合表征学习的开放环境下作物病害识别模型(bimodalNet)。该模型在病害图像模态的基础上,进行了病害文本模态信息的嵌入,利用两种模态病害信息间的相关性和互补性,实现了病害特征的联合表征学习。最终bimodalNet在较小的数据集上取得了优于单纯的图像模态模型和文本模态模型的效果,最优模型组合在测试集的准确率、精确率、灵敏度、特异性和F1值分别为99.47%、98.51%、98.61%、99.68%和98.51%。该研究证明了利用病害图像和病害文本的双模态表征学习是解决开放环境下作物病害识别的有效方法。     

基于迁移学习的GoogLeNet果园病虫害识别与分级

为提高果树病虫害危害程度分级精度进而更好地指导果园病虫害防治，采用迁移学习技术与GoogLeNet模型相结合的方法，对6种果园作物的25类病虫害样本进行识别与危害程度分级研究；同时，探究不同数据集大小以及不同优化算法对模型性能的影响；基于MATLAB平台设计了一款可视化的病虫害识别与分级系统。结果表明:1)基于迁移学习的GoogLeNet模型，对病虫害识别精度可达99.35%，危害程度分级精度可达92.78%；2)在相同训练参数下，本研究模型比AlexNet、VGG-16、ResNet-18、SqueezeNet、原GoogLeNet及MobileNet-v2模型验证精度提高了2.38%~11.44%，并且收敛速度最快；3)本研究模型识别精度随着数据集的增大而提高；在3种优化算法中SGDM算法耗时最短且精度最高，更适合本研究模型。通过拍摄果树叶片病害区域图像，本研究设计的系统能够在0.43 s左右准确识别出果树种类、病害类型以及危害等级等信息。     

油料作物油脂合成调控研究进展

植物油脂不仅是人类可食用油的主要来源,也是人类生产生活中重要的可再生原料。本文概述了植物油脂的生物合成途径,从母体效应、QTL、GWAS等多个方面总结了油料作物油脂合成的遗传学研究进展,同时探讨了已知的油脂合成调控相关基因的功能。本文综述了该领域的研究现状,为深入了解油料作物油脂合成调控网络提供了参考,也为油料作物的分子改良和遗传育种提供了理论基础。     

茶树植保机械及减量施药技术研究进展

农药防治是茶树病虫害综合防治的重要组成部分,其在病虫害突发或爆发时具有快速高效的防治优势。茶树叶片表面具有亲水性,常量施药会造成茶叶农残超标、生态环境破坏等问题,实现茶树减量施药是减少茶叶农残的有效手段。系统综述了茶树生物特性、茶树病虫害预测诊断及防治方法、茶树植保机械及施药技术,强调提高茶树低容量喷雾的农药有效利用率是实现茶树减量施药的关键。针对目前茶园地面工况复杂及农药利用率低的问题,本文从低容量仿形喷雾机、茶树病虫害喷雾决策及智能终端等六个方面提出茶树病虫害施药技术及装备的研究建议,指出低量化、精准化及智能化是未来茶树植保喷雾机械及施药技术的发展方向。     

南充地区晚熟柑桔主要病虫害发生特点及综合防控措施

本文介绍了南充地区晚熟柑桔主要病虫害的种类及为害发生特点,并提出了相应的病虫害综合防控措施,保证晚熟柑桔的正常生长,以期为当地晚熟柑桔生产、种植户节本增收提供参考。     

乡村振兴战略背景下江西省设施蔬菜发展问题与对策

为加快乡村振兴战略背景下江西省设施蔬菜产业发展,实现设施蔬菜增产增效,以江西省乐平、永丰、高安、崇仁、瑞金和樟树等6个农业大县(市)为例,分析了江西省设施蔬菜的发展现状和存在问题,并提出了科学把握发展方向、积极壮大投资主体等5个方面的对策建议,以期为当地设施蔬菜发展提供新的依据与参考。     

北京地区南瓜病毒病病原种类鉴定

为明确北京地区南瓜病毒病种类及其主要侵染病原,2016~2017年在北京周边采集疑似感染病毒的南瓜病样84份,并根据南瓜上的6种病毒特异性引物对其进行反转录PCR(RT-PCR)检测。结果表明:共有79份南瓜病样检测显示阳性,其中黄瓜花叶病毒(Cucumber mosaic virus,CMV)的检出率最高,为52.38%;其次是小西葫芦黄花叶病毒(Zucchini yellow mosaic virus,ZYMV)和西瓜花叶病毒(Watermelon mosaic virus,WMV),检出率分别为44.01%、14.29%,其他病毒暂未检出。此外,16.67%的样品受2种病毒复合侵染,CMV和ZYMV复合侵染占7.14%,ZYMV和WMV复合侵染占9.52%。北京地区南瓜上优势病毒种类为CMV,且存在病毒复合侵染现象。     

花椒病虫害研究进展

花椒Zanthoxylum bungeanum是一种具有重要经济价值和药用价值的多年生香料和油料树种,国内外学者对花椒的栽培、育种等方面研究较多,而对花椒病虫害防治方面的研究较少。笔者在前人研究的基础上,通过文献分析明确花椒及其病虫害方面的研究,并对已报道的花椒病虫害、防治方法等进行了综述,以期为今后开展花椒病虫害防治相关研究提供理论支持和方向。     

Prevention and Controlling Effect of Intercropping on Pests and Diseases and Its Risk Control in Cotton Fields

Cotton is one of the major field crops which are seriously threatened by pests and diseases. In recent years, the ecological management of cotton pests and diseases with intercropping has become to be an interesting approach. Intercropping possibly increases the population of natural enemies and finally reduces the population density of pests as a result of changes in ecological structure and environmental conditions of farmland. Intercropping thus has been considered an important alternative in controlling pests and diseases. However, cotton field intercropping also has some limitations in controlling pests and diseases. Unreasonable intercropping system has many risks such as increasing labor input, increasing the difficulty of pests’ control, and aggravating pests and diseases. In this paper, the effects and the underlying mechanisms of cotton intercropping on cotton pests and diseases were reviewed. The possible risks and countermeasures used for ecological control of intercropping were also put forward, and the future and application of ecological management of cotton pests and diseases with intercropping were prospected.     

Intensive organic vegetable production increases soil organic carbon but with a lower carbon conversion efficiency than integrated management

Intensive vegetable production in greenhouses has rapidly expanded in China since the 1990s and increased to 1.3 million ha of farmland by 2016, which is the highest in the world. We conducted an 11‐year greenhouse vegetable production experiment from 2002 to 2013 to observe soil organic carbon (SOC) dynamics under three management systems, i.e., conventional (CON), integrated (ING), and intensive organic (ORG) farming. Soil samples (0–20 and 20–40 cm depth) were collected in 2002 and 2013 and separated into four particle‐size fractions, i.e., coarse sand (> 250 µm), fine sand (250–53 µm), silt (53–2 µm), and clay (< 2 µm). The SOC contents and δ¹³C values of the whole soil and the four particle‐size fractions were analyzed. After 11 years of vegetable farming, ORG and ING significantly increased SOC stocks (0–20 cm) by 4008 ± 36.6 and 2880 ± 365 kg C ha^?1 y^?1, respectively, 8.1‐ and 5.8‐times that of CON (494 ± 42.6 kg C ha^?1 y^?1). The SOC stock increase in ORG at 20–40 cm depth was 245 ± 66.4 kg C ha^?1 y^?1, significantly higher than in ING (66 ± 13.4 kg C ha^?1 y^?1) and CON (109 ± 44.8 kg C ha^?1 y^?1). Analyses of ¹³C revealed a significant increase in newly produced SOC in both soil layers in ORG. However, the carbon conversion efficiency (CE: increased organic carbon in soil divided by organic carbon input) was lower in ORG (14.4%–21.7%) than in ING (18.2%–27.4%). Among the four particle‐sizes in the 0–20 cm layer, the silt fraction exhibited the largest proportion of increase in SOC content (57.8% and 55.4% of the SOC increase in ORG and ING, respectively). A similar trend was detected in the 20–40 cm soil layer. Over all, intensive organic (ORG) vegetable production increases soil organic carbon but with a lower carbon conversion efficiency than integrated (ING) management.