日本设施农业采收机器人研究应用进展及对中国的启示

设施农业智能装备是设施农业稳定、高品质、高效生产的必要保障。日本智能采收装备已有近四十年的研发经验,其发展具有一定启发和借鉴意义。本文综述了日本设施农业采收机器人的研究应用进展,分析了基于农机农艺结合的茄科（番茄、茄子、青椒）、葫芦科（黄瓜、瓜类水果）、芦笋和草莓等10种设施农业采收机器人的采收技术,其中详细对比了番茄、草莓等几种蔬菜历代采收机器人的设计理念及其优点与不足。分析了设施农业采收机器人面临的科学问题及解决方案,总结了未来发展趋势及对中国的启发。本文可为加速推进中国设施农业采收机器人的智慧化、智能化和产业化发展提供借鉴参考。     

Does sika deer overabundance exert cascading effects on the raccoon dog population?

Habitat structure in Oku-Nikko, Japan, has been substantially modified by the overabundance of sika deer (Cervus nippon). A deer-proof fence (15.0 km and 900 ha) was constructed in 2001 to conserve vegetation. Although the understory inside the fence is dominated by Sasa nipponica (hereafter, Sasa), an important forage plant for deer, that outside the fence is dominated by Aster ageratoides leiophyllus (hereafter, Aster), an unpalatable plant to deer, and, partly, by bare floor. In this study, we examined the effects of deer on ground-dwelling insects and earthworms, the primary food resources of raccoon dogs (Nyctereutes procyonoides), and, thus, the bottom-up cascading effects of the herbivore on the omnivorous carnivore. Between July and September 2008, we examined the abundance of insects and earthworms by pitfall trapping and hand-sorting methods, respectively, both inside and outside the fence. The abundance of earthworms and insects (Scarabaeidae and Rhaphidophoridae) was higher on forest floors with Aster and/or bare floors outside the fence than on those with Sasa inside the fence. These results indicate that the increasing deer population in this area probably increased the number of these invertebrates outside the fence by modifying understory vegetation and/or depositing dung. Furthermore, the sighting rates of raccoon dogs obtained by spotlight counts were greater outside than inside the fence, suggesting that deer probably exert bottom-up cascading effects on raccoon dogs, at least during May to November, when the invertebrates are predominantly fed on by the omnivorous carnivore.     

Evaluation of the PATHFAST Chemiluminescent Enzyme Immunoassay for Measuring Progesterone in Whole Blood and Serum of Mares

Evaluation of a new chemiluminescent enzyme immunoassay, the PATHFAST assay system (PATHFAST), for measurement of circulating progesterone in mares was performed. Five mares at the mid-luteal stage were administrated a single i.m. injection of prostaglandin F2α analog (PGF2α; cloprostenol 250 μg/ml), and then blood samples were collected from the jugular vein at 0, 15, 30 and 45 min, at one-hour intervals until 24 and at 48 hr via a catheter in the jugular vein. To monitor the physiological changes in circulating progesterone in mares after induced luteolysis, concentrations of progesterone in whole blood and serum samples were measured by PATHFAST. In addition, concentrations of progesterone in serum samples measured by PATHFAST were compared with those measured by radioimmunoassay (RIA) and enzyme immunoassay (EIA). Using PATHFAST, the serum concentrations of progesterone in mares correlated highly with those of whole blood samples (r=0.9672, n=88). The serum concentrations of progesterone as measured by PATHFAST correlated well with RIA (r=0.9654, n=88) and EIA (r=0.9323, n=112). An abrupt decline in circulating progesterone in whole blood samples was observed within 2 hr (50%), followed by a gradual decline until 48 hr later. The results for progesterone in whole blood samples correlated highly with those in serum samples, and the declining pattern paralleled that of the serum samples. These results demonstrated that PATHFAST is useful in the equine clinic as an accurate diagnostic tool for rapid assay of progesterone within 26 min, using unextracted whole blood.     

Phenotypic changes in reproductive traits with changes in stock size of the starspotted smooth-hound Mustelus manazo in Tokyo Bay, Japan

Reproductive traits are among the primary determinants of population dynamics, and their changes could affect stock size. We investigated differences in reproductive traits of the starspotted smooth-hound Mustelus manazo in Tokyo Bay between the 1990s (a low-stock period) and the 2000s (a high-stock period). Peak timings of mating (May–July), ovulation (May–June), and parturition (May–June) were similar between the two periods. However, the size at first maturity in the 2000s was smaller than that in the 1990s for both sexes. In addition, embryo-related parameters differed greatly between the two periods. Slopes of size–fecundity regression in the 2000s decreased to 43–73 % of the 1990–1996 values, suggesting a substantial decline in fecundity during the 2000s. Frequency distribution of the maximum ovum diameter during the ovulation season showed two distinct two modes in 1990s while it showed a substantial dispersion during the 2000s, implying that changes occurred in the developmental pattern of ova in the ovary. Undeveloped eggs were present in the uteri with low frequency from September to May during the 2000s, but they were totally absent from September through May in the 1990s. The increased number of remnant undeveloped eggs that did not become embryos may explain the fecundity decline during the 2000s.     

Estimation of annual suspended sediment yield from a Japanese cypress (Chamaecyparis obtusa) plantation considering antecedent rainfalls

Japanese cypress (Chamaecyparis obtusa) is one of the most common plantation species in Japan. In most Japanese cypress plantations, closure of the forest canopy hinders understory growth because of insufficient management practices. Thus leaf litter barely covers the soil surface. Such plantations are vulnerable to surface erosion triggered by rainfalls, and could yield large amounts of suspended sediment (SS). However, few studies have investigated the annual SS yield. This study aimed to develop a modified model of SS yields, and to accurately estimate and characterize the annual SS yield from a Japanese cypress plantation. For this, hydrological surveys were conducted for 5 years, and stream water was sampled weekly as well as sequentially at 15–60 min intervals in each of 14 rain events in a small forested watershed of Japanese cypress. The root mean square error (RMSE) and the annual SS yield estimates were obtained using three different equations: the rating curve derived from instantaneous SS concentration versus stream discharge relationship (SSC − Q equation); the rating curve derived from cumulative SS yield versus cumulative specific discharge relationship (∑SS − ∑Q equation); and the multiple regression of cumulative SS yield expressed as functions of cumulative specific discharge and antecedent rainfalls (∑SS − ∑Q × AR equation). The ∑SS − ∑Q × AR equation is a modified model considering the effects of antecedent rain history on the SS yield. The parameters in each equation were determined by a nonlinear least square method. Of the three equations, the RMSE was the highest using the SSC − Q equation. Although the ∑SS − ∑Q equation and the ∑SS − ∑Q × AR equation gave similar RMSE values, the SS yields extrapolated by the ∑SS − ∑Q equation would be overestimated. Thus, the 5-year average of annual SS yield estimated by the ∑SS − ∑Q equation was three times larger than that estimated by the ∑SS − ∑Q × AR equation. These results indicate that the ∑SS − ∑Q × AR equation more accurately estimates the annual SS yield, compared with the other two equations. The results also suggest that antecedent rain history, as well as stream discharge, should be considered for appropriately estimating the annual SS yield. The annual SS yield estimated by the ∑SS − ∑Q × AR equation was the largest among 25 undisturbed forested watersheds, except for two watersheds dominated by sandstone. Our results suggest that the annual SS yield is greater from a Japanese cypress plantation under current management practices than from other undisturbed forests.     

Yield-enhancing and tuber-downsizing effects of transplantation cultivation method of case-held tuber seedlings in the sweet potato cultivar Beniharuka

We developed transplantation cultivation method of case-held tuber seedlings (CTS), which was derived from direct planting method of seed tubers, and applied this method to the sweet potato cultivar Beniharuka. A plastic case made of polypropylene was designed for cultivation of CTS. Seed tubers of cultivar Beniharuka in the range of 30–80 g were cut in half. The half-cut tubers were placed inside the plastic cases, and the cases were filled with a commercial soil mix. The case-held tubers were incubated under natural sunlight in a glass house. After 3–4 wk, the CTS were transplanted into a field. Mother tuber (seed tuber) enlargement was suppressed by the plastic confinement of the cases, and daughter tubers were formed above the case as vine-root-originated tubers. In the field experiments in 2012 and 2013, daughter tuber yields were increased 19% and 21% by case-held tuber seedling transplanting (CTST) over conventional vine-planting (VP), the number of daughter tubers per plant in CTST were 36 and 68% higher than in VP, and the mother tuber yields were limited to 2.1 and 4.3% of the total fresh yield of mother and daughter tubers, respectively in 2012 and 2013. Application of CTST method to cultivar Beniharuka enhanced tuber yield, increased the number of daughter tubers per plant, downsized daughter tubers compared to VP, and mother tuber enlargement was suppressed by case-holding. The CTST method is expected to produce more and smaller good in shape tubers of cultivar Beniharuka compared to VP.     

An application of Neurogenetic Algorithm System to individual tree growth model

A neural network system with genetic algorithms (Neurogenetic Algorithm System, or NGAS) was employed to develop individual coniferous tree growth models. A multivariable regression model was applied to compare the performance of NGAS. An IBM personal computer with the BioComp System’s software program of NGO was used to execute this comparison. The results indicate that NGAS is more accurate and effective than the conventional regression method in modeling individual tree growth based on the criteria of Sum of Squared Error (SSE), Average Absolute Error (AAE), Mean Squared Error (MSE) and Final Predicted Error (FPE). This study also suggests that individual tree growth may indeed be a non-linear process. Using this flexible neural network system to model individual tree growth can yield satisfactory prediction results.