基于计算机视觉和土壤Nmin的棉田氮素养分管理系统的建立与验证

Construction of Cotton Field N Management System Based on Computer Vision and Soil Nmin

【目的】报道一种基于计算机视觉和土壤(Nmin)的棉花氮素营养诊断和氮肥追肥推荐软件FertiExp系统的建立及其示范应用。【方法】通过棉花需氮量吸收模型、棉花吸氮量计算机视觉识别模型和土壤无机氮供应估算模型建立了FertiEXP软件系统。【结果】小区试验结果表明:采用FertiEXP进行氮肥追肥管理,氮肥利用率由37.85%提高到了56.02%,相对提高48.0%,产量基本持平。FetiEXP氮肥土壤中的残留量比常规管理减少100 kg/hm2以上,氮素的表观损失率也大幅度降低。大田示范结果表明,采用FertiEXP推荐施肥,可以优化肥料在作物全生育期的分配比例,施N量降低6.7%～14.6%,而产量提高5.8%～7.5%。【结论】采用FertiEXP推荐施肥经济效益十分显著。

【Objective】This paper focused on reporting cotton field N nutrient management of computer rision and soil Nmin and construction of recommending dressing N fertilizer of software FertiEXP system and its utilization. 【Method】FertiEXP software system was constructed based on cotton N nutrient absorption model,cotton N nutrient absorption model designed by computer program and estimation model of soil inorganic N use efficiency.【Result】Plot trails demonstrated that FertiEXP system could promote N use efficiency from 37.85% to 56.02%,with a relatively rise by 48.0% while harvested Equivalent yield.N residue in soil reduced by 100 kg/hm2 by FertiEXP as compared with farmers conventional practice,apparent N lose was reduced significantly.Field demonstration showed that FertiEXP could optimal N distribution in the cotton＇s whole growth period,N rate reduced by 6.7% to 14.6%,yield increased by 5.8% to 7.5%.【Conclusion】FertiEXP produced a sound economical profile.