基于植物工厂LED阵列补光灯的栽培槽平面光场强度预测模型.

【Objective】According to the climatic characteristics such as the lack of sunshine in Chongqing, combined with the different light demand of different crops and different growth stages, the present study explored the distribution law of the plane light field of the cultivation tank under the conditions of different plant fill light heights and light intensity ratio in the plant factory, and established a quantitative relationship model between the fill light height, light intensity ratio and the plane light field intensity of the cultivation tank, so as to provide support for the automatic, accurate and on-demand regulation of the fill light height and light intensity. 【Method】In the closed intelligent simulation agricultural environment laboratory of Chongqing Academy of Agricultural Sciences, and taking the intelligent LED fill light as the research object, the paper explored the light field intensity change law of the cultivation tank under the conditions of different fill light height and light intensity ratio. The regression support vector machine algorithm was used to construct the spatial light field intensity prediction model of fill light, which took the fill light height, fill light intensity ratio and spatial position information as the input and the plane light field intensity of cultivation tank as the output. Based on this model, the intensity of the plane light field in the cultivation tank was predicted, and the height of the fill light suitable for different crops and different growth stages was adjusted as needed. 【Result】Exploring the law of light field distribution on the plane of cultivation tank under different heights and light intensity ratios, it was found that with the increase of the proportion of light intensity at the same height, the value of light field intensity on the surface of the cultivation tank plane light field also increased; Under the same light intensity ratio, with the increase of fill light height, the light field intensity value in the center area of the plane light field surface of the cultivation tank decreased, while the light field intensity value in the edge area of the light field surface increased to a certain extent. By comparing the five modeling methods, it was found that SVR algorithm had the best performance on the data set. At the same time, the training accuracy R~2 of the plane light field intensity prediction model of the cultivation tank constructed in this study could reach 0.9897, and the testing accuracy R~2 could reach 0.9815, which proved that the model had good reliability. 【Conclusion】 The distribution law of light field intensity in the cultivation tank plane of LED array fill light in plant factory was analyzed, which proved the necessity of building the prediction model of light field intensity in the cultivation trough plane. The precise construction of this model has laid a good support for the optimal dynamic regulation of light source location and intensity in different crops and different growth stages under the conditions of installment agriculture or plant factories. [ABSTRACT FROM AUTHOR]