基于级联卷积神经网络的番茄花期识别检测方法

对作物花期状态的准确识别是温室作物执行授粉的前提。为提高花期状态识别的准确度，该研究以温室番茄为例提出了一种基于级联卷积神经网络的番茄花朵花期识别方法。首先采用改进的端到端的特征金字塔网络FS-FPN实现番茄花束的分割，然后采用Prim最小生成树对分割后的花束图片进行花期识别优先级排序，最后将已排序的分割花束图片输入改进的Yolov3网络，实现番茄花朵花期状态的精准识别。在由1600幅包含花蕾期、全开期、谢花期、初果期4类花期状态的番茄花束图像构成的数据集上，所提方法对番茄花朵花期平均检测时间为12.54 ms，平均检测精度分别比Mask R-CNN和SPP-Net提高了3.67%和2.39%，识别错误率比改进前的Yolov3网络降低了1.25%。最终将该方法部署到番茄授粉机器人上，并在大型玻璃温室内进行验证，结果表明，所提方法对番茄花朵各花期的检测精度分别为花蕾期85.71%、全开期95.46%、谢花期62.66%、初果期88.34%，该研究结果可为智能授粉机器人的精准作业提供重要依据。     

Simulation of future dissolved oxygen distribution in pond culture based on sliding window-temporal convolutional network and trend surface analysis

Dissolved oxygen (DO) is a key ecological factor to measure the quality of water in the aquaculture. As the pond water body is affected by the breeding environment, the spatial distribution of DO shows a certain law in the entire pond. Therefore, to simulate the distribution of DO in aquaculture waters and grasp the temporal and spatial variation of DO is the key to achieving precise regulation of DO. For this purpose, this paper proposed a method for simulating the temporal and spatial distribution of DO in pond culture based on a sliding window-temporal convolutional network together with trend surface analysis (SW-TCN-TSA). This paper first utilized SW to construct DO data sets with different prediction durations, and then used the improved TCN model to realize one-dimensional time series prediction for DO at single monitoring point. Based on the prediction results of DO, a TSA method was performed on the predicted values of DO at the extreme moments of all discrete monitoring points, so as to realize the simulation of the temporal and spatial distribution of DO in the pond. Experimental results show that the SW-TCN model has better prediction performance for one-dimensional time series prediction of DO. Compared with traditional deep networks, such as CNN, GRU, LSTM, CNN-GRU and CNN-LSTM, the values of evaluation indicators (MSE, MAE and RMSE) have been greatly improved. In the process of trend surface fitting, all fitting R² of DO at different water depths are higher than 0.9, indicating that the TSA can accurately reflect the temporal and spatial distribution of DO. This method can provide a basis for the prediction and early warning of DO in the three-dimensional space of the pond and has high practicability in aquaculture.     

基于农作制分区的1985—2015年中国小麦生产时空变化

Comparing spatio-temporal variation characteristics of China’s wheat production, yield, sown area and yield-area-contribution in different farming zones in the past 30 years could help improving wheat planting layout and adjusting planting structure. Concentration index, rate of change, moving of gravity and resolution of yield-area-contribution were used to analyze spatio-temporal changes of China’s wheat production and yield-area-contribution based on county wheat production statistics including sown area, production and yield from 1985 to 2015. The wheat sown area decreased obviously in Northeast farming region, Northwest farming region and South farming region and increased rapidly in Huang-Huai-Hai farming region and Yangtze Plain farming region. In Huang-Huai-Hai farming region, the concentration indexes of Haihe Plain farming region, Huang-huai Plain farming region and Fenwei Basin farming region reached to 20.64%, 25.77%, and 21.65% respectively in 2015. Wheat production increased significantly by more than 48 milliontons in Huang-Huai-Hai farming region, and nearly 8 million tons Yangtze Plain farming region but decreased by more than 2.6 million tons in Northeast farming region. In Huang-Huai-Hai farming region, wheat production was concentrated in Haihe Plain farming region, Huang-huai Plain farming region and Yuxi Hill farming region. The average wheat yield continuously improved during the study period, Huang-Huai-Hai farming region and Northwest farming region had the yield increase up to 103.5 kg hm ^-2 and 92.9 kg hm ^-2 each year. Wheat yield in Yuxi Hill farming region, Fenwei Basin farming region and Haihe Plain farming region was relatively high in Huang-Huai-Hai farming region. The yield-reduced area was mainly caused by the decreasing of sown area, while the yield-area-contribution rate was different in yield-increased area. Yield-dominant counties were reduced, area-dominant counties were increased and yield-area-dominant counties were relatively steady. In production increased area, yield-dominant and yield-area-dominant counties were the main types in Huang-Huai-Hai Farming region, area-dominant and yield-area-dominant counties were the main types in Yangtze Plain farming region. Chinese wheat production was increasingly concentrated in Huang-Huai-Hai farming region which has high and rapid increase of wheat yield over the past three decades. Haihe Plain farming region, Huang-huai Plain farming region and Fenwei Basin farming region were the most concentrated areas in Huang-Huai-Hai farming region for wheat production during this period. Wheat yield and sown area jointly promoted the increase of wheat production in Huang-Huai-Hai farming region, wheat sown area was the crucial factor to increase wheat production in Yangtze Plain farming region, especially in the north of Jiangsu, Anhui province and greater part of Xinjiang.     

用于冬小麦组分信息检测的便携式LVF光谱仪研究

小麦是我国主要的粮食作物,能够实时准确地对小麦进行长势监测具有重要意义。为此,介绍了一种自主研发的便携式LVF光谱仪的软硬件结构,采集冬小麦冠层叶片光谱数据,利用偏最小二乘法建立了冬小麦组分信息的内嵌反演模型,使其成为可以测量冬小麦叶片组分信息的专用型光谱仪。同时,用校验集对预测模型进行验证,相关系数分别达到0.906和0.876。结果表明:便携式LVF光谱仪在冬小麦叶片组分信息的检测方面是可行性的。     

Monitoring wheat phenology and irrigation in Central Morocco: On the use of relationships between evapotranspiration,crops coefficients,leaf area index and remotely-sensed vegetation indices

The monitoring of crop production and irrigation at a regional scale can be based on the use of ecosystem process models and remote sensing data. The former simulate the time courses of the main biophysical variables which affect crop photosynthesis and water consumption at a fine time step (hourly or daily); the latter allows to provide the spatial distribution of these variables over a region of interest at a time span from 10 days to a month. In this context, this study investigates the feasibility of using the normalised difference vegetation index (NDVI) derived from remote sensing data to provide indirect estimates of: (1) the leaf area index (LAI), which is a key-variable of many crop process models; and (2) crop coefficients, which represent the ratio of actual (AET) to reference (ET₀) evapotranspiration.A first analysis is performed based on a dataset collected at field in an irrigated area of the Haouz plain (region of Marrakesh, Central Morocco) during the 2002–2003 agricultural season. The seasonal courses of NDVI, LAI, AET and ET₀ have been compared, then crop coefficients have been calculated using a method that allows roughly to separate soil evaporation from plant transpiration. This allows to compute the crop basal coefficient (K_cb) restricted to the plant transpiration process. Finally, three relationships have been established. The relationships between LAI and NDVI as well as between LAI and K_cb were found both exponential, with associated errors of 30% and 15%, respectively. Because the NDVI saturates at high LAI values (>4), the use of remotely-sensed data results in poor accuracy of LAI estimates for well-developed canopies. However, this inaccuracy was not found critical for transpiration estimates since AET appears limited to ET₀ for well-developed canopies. As a consequence, the relationship between NDVI and K_cb was found linear and of good accuracy (15%).Based on these relationships, maps of LAI and transpiration requirements have been derived from two Landsat7-ETM+ images acquired at the beginning and the middle of the agricultural season. These maps show the space and time variability in crop development and water requirements over a 3 km × 3 km irrigated area that surrounds the fields of study. They may give an indication on how the water should be distributed over the area of interest in order to improve the efficiency of irrigation. The availability, in the near future, of Earth Observation Systems designed to provide both high spatial resolution (10 m) and frequent revisit (day) would make it feasible to set up such approaches for the operational monitoring of crop phenology and irrigation at a regional scale.     

长时序海量土地利用时空数据统计优化方法

针对任意查询区域年度现状地类面积统计困难、长时序变更流量分析计算耗时等问题,提出基于时空变化图模型的统计优化方法。运用图的连通性原理,对查询统计区域内和边界处的要素实体进行分类,实现了时序快照统计优化算法,解决了任意查询区域时点现状统计困难的问题,提高了时序快照统计的效率。运用多商品流原理进行时空网络图约化性判定,实现了变更流量统计优化算法,减少了要素空间叠置分析次数,解决了长时序土地利用变化变更流量统计耗时问题,提高了统计的效率。最后,以2009—2012年琼海市土地利用数据为例,进一步验证优化算法的有效性和可行性。     

Yield performance of spring wheat improved by regulated deficit irrigation in an arid area

A field experiment was conducted in 2003 and 2004 growing seasons to evaluate the effects of regulated deficit irrigation on yield performance in spring wheat (Triticum aestivum) in an arid area. Three regulated deficit irrigation treatments designed to subject the crops to various degrees of soil water deficit at different stages of crop development and a no-soil-water-deficit control was established. Soil moisture was measured gravimetrically in the increment of 0–20 cm every five to seven days in the given growth periods, while that in 20 increments to 40, 40–60, 60–80, and 80–100 cm depth measured by neutron probe. Compared to the no-soil-water-deficit treatment, grain yield, biomass, harvest index, water use efficiency (WUE), and water supply use efficiency (WsUE) in spring wheat were all greatly improved by 16.6–25.0, 12.4–19.2, 23.5–27.3, 32.7–39.9, and 44.6–58.8% under regulated deficit irrigation, and better yield components such as thousand-grain weight, grain weight per spike, number of grain, length of spike, and fertile spikelet number were also obtained, but irrigation water was substantially decreased by 14.0–22.9%. The patterns of soil moisture were similar in the regulated deficit treatments, and the soil moisture contents were greatly decreased by regulated deficit irrigation during wheat growing seasons. Significant differences were found between the no-soil-water-deficit treatment and the regulated soil water deficit treatments in grain yield, yield components, biomass, harvest index, WUE, and WsUE, but no significant differences occurred within the regulated soil water deficit treatments. Yield performance proved that regulated deficit irrigation treatment subjected to medium soil water deficit both during the middle vegetative stage (jointing) and the late reproductive stages (filling and maturity or filling) while subjected to no-soil-water-deficit both during the late vegetative stage (booting) and the early reproductive stage (heading) (MNNM) had the highest yield increase of 25.0 and 14.0% of significant water-saving, therefore, the optimum controlled soil water deficit levels in this study should range 50–60% of field water capacity (FWC) at the middle vegetative growth period (jointing), and 65–70% of FWC at both of the late vegetative period (booting) and early reproductive period (heading) followed by 50–60% of FWC at the late reproductive periods (the end of filling or filling and maturity) in treatment MNNM, with the corresponding optimum total irrigation water of 338 mm. In addition, the relationships among grain yield, biomass, and harvest index, the relationship between grain yield and WUE, WsUE, and the relationship between harvest index and WUE, WsUE under regulated deficit irrigation were also estimated through linear or non-linear regression models, which indicate that the highest grain yield was associated with the maximum biomass, harvest index, and water supply use efficiency, but not with the highest water use efficiency, which was reached by appropriate controlling soil moisture content and water consumption. The relations also indicate that the harvest index was associated with the maximum biomass and water supply use efficiency, but not with the highest water use efficiency.     

精密播种机下压力和播深CAN总线监控与评价系统研究

为实现精密播种作业中播种下压力和播深的实时监控和质量评价,设计了一种多行播种机下压力和播深CAN总线监控与评价系统。系统采用基于角度和轴销传感器的播深和下压力测量装置,优化设计了液压驱动和分区控制的气压驱动装置,开发了基于Co De Sys(Controlled development system)编程环境的智能终端交互界面和ECU(Electronic control unit)控制程序,实现了基于CAN总线通信的作业参数监测控制和质量评价。通过搭建的室内试验台完成了播深和下压力静态建模试验,建立了适应不同设定播深的下压力测量模型。分区控制系统响应测试试验表明,在调节范围(0. 2～0. 6 MPa)内,系统超调量低于5. 97%;响应时间与控制行数和设定气压正相关;在设定气压(0. 1～0. 6 MPa)范围内,6行播种机调节时间不超过2. 35 s。为测试系统工作性能,在25、50、75 mm 3种设定播深下,对左区控制(600 N)、右区控制(300 N)、机械调节和自重调节4种控制方式进行了田间性能试验。土壤压实和播种下压力控制效果试验表明,主动分区控制方式可实现更为稳定的土壤紧实度,且在浅旋地块环境下,右区控制方式可达到最优的下压力稳定性,其控制合格率不小于95. 78%;播深控制效果试验表明,随着设定播深的增大,播深质量显著降低,在设定播深25～75 mm范围内,左区控制、右区控制、机械调节和自重调节对应的最小播深合格率分别为91. 92%、92. 53%、70. 44%和58. 72%,对应的最大标准差分别为2. 22、3. 11、3. 69、7. 70 mm,对应的最大变异系数分别为3. 52%、4. 40%、4. 96%和14. 01%。相比机械调节和自重调节,分区控制系统提高了单体下压力和播深稳定性。     

A longitudinal study of the effect of time on the matching performance of a retinal recognition system for lambs

While there is strong evidence supporting retinal vascular pattern as a distinctive marker for sheep, it would be advantageous to get an insight into its robustness; in other words, to determine whether retinal recognition of young animals (lambs) can reach as good a matching performance as the one demonstrated for adult sheep. To this aim, a longitudinal study was devised to observe the evolution of matching scores (ms) of lamb retinal images (n = 38) acquired from 1 to 22 weeks after birth. It was observed that four lamb retinas (out of 38) underwent slight curving of one or two secondary arteries, which ceased by the time they were 6–8 weeks old. However, this slight artery curving did not affect matching performance. A random effects statistical model demonstrated that lamb age had an effect (P < 0.01) on the matching scores produced using this commercially available retinal recognition system. As lambs grew older (larger eyes) and they became easier to handle, retinal images of progressively better quality could be obtained in a more consistent way; and thus matching scores increased from an average of 86 at the first week of life, up to an average of 96 by week 8. Dunnett simultaneous tests of means indicated that no further improvement in matching score took place once lambs were at least 6–8 weeks old, meaning that the retinal image quality became by then optimal and consistent. Although the variable retinal image quality of younger lambs (1–4 weeks old) caused a reduction in matching score, they did not lead to false non-matches in any case (considering a cut-off matching score of 70 for acceptance of a positive match). Therefore, the results of these trials have shown that, with the available technology, retinal images can be used as a robust biometric marker of lambs from 1 week of age.     

云计算在我国农业信息服务中的研究现状与思考

利用云计算技术进行农业信息服务是当前信息农业研究的热点。分析了云计算技术在农业科研创新、涉农用户使用以及基层农技推广中应用的优势，综述了农业信息云服务的基础架构、云环境资源管理、云环境数据挖掘、云服务应用的国内研究现状，最后针对当前研究存在的问题提出了对策建议，以便为利用云计算创新农业信息服务研究提供有益的参考。