Introduction to the project

With regard to the use of procedures based on spectrophotometry and hyperspectral sensors, architectures (i.e. portable hardware and software integrated systems) operating both at a point scale and at a crop image scale will be developed, implemented and fine-tuned, all this in order to be able to evaluate in an almost continuous manner the presence/absence of chlorates and perchlorates in crops (i.e. lamb's lettuce), in relation to possible different methods of cultivation in greenhouses. These procedures will also be used in the post-harvest phases. The identification and recognition of specific plant characteristics, as well as their classification based on specific attributes (i.e. absence of chlorates and perchlorates), represent important aspects, which if analyzed through easy-to-use, low-cost methods and above all not requiring sample collection and preparation, represent a fundamentally important tool in order to operate an almost continuous control and monitoring of both the plant (i.e. valerian), as well as the characteristics of the soil, with specific reference to the conditions of cultivation in greenhouses and the management of fertilizers (i.e. based on chlorates and nitrates) in relation to specific actions and/or environmental conditions. To this end, spectroscopic and/or hyperspectral investigation procedures will be developed starting from the analysis of the plants, the characteristics and the state of the soil. The innovative spectroscopic procedures will then be compared, for each of the investigated typologies (i.e. plants and soils), with the classical analytical ones, currently used, all this in order to build correlation models/procedures, through which it will then be possible to determine the attributes of interest of the different typologies (i.e. plants, and soils) starting from the sole survey of the directly acquired hyperspectral attributes. The project will be able to take place thanks to the formation of a partnership. This will be composed of agricultural and marketing companies belonging to the fruit and vegetable sector of the Lazio region and in particular of the Agro Pontino area whose objective will be to obtain a specific product: valerianella with zero chlorate and perchlorate content. A public research body, the Research and Services Center for Sustainable Technological Innovation (Ce.R.S.I.Te.S) of the University of Rome – La Sapienza, and two private research bodies will also be involved: Eurolab S.r.l. and the European Center for Managerial Studies. Other professional figures will be involved in this project. The research activities will be carried out through: • the use of innovative non-destructive and non-invasive monitoring techniques aimed at evaluating the effects of the application of a protocol that guarantees the production of valerianella with low nitrate content through the study of the various changes recorded during cultivation; • the development, fine-tuning, implementation and verification of techniques based on hyperspectral digital spectrophotometry operating in the VIS (400-1000 nm), NIR (1000-1700 nm) and SWIR (1000-2500 nm) ranges, capable of carrying out rapid, simple and efficient monitoring of the state of plants and soils in relation to different cultivation protocols. The methodologies used will be those specific to chemometric analyses and recognition and classification algorithms currently used for this type of investigation. To achieve this objective, from a software point of view, we will use the libraries contained in the PLS_Toolbox (Eigenvector Research, Inc.) operating in a MATLAB environment (The Mathworks, Inc.). Specific procedures developed autonomously and originally by the research group will therefore be developed, implemented, fine-tuned and validated. The investigations will be developed through acquisition, archiving, pre-processing and analysis of hyperspectral data (i.e. digital spectra of plants and soils from 400-2500 nm at different spectral resolution). This approach will be aimed at developing classification/recognition models of both the different state of valerianella plants in relation to the content of chlorates and perchlorates in the leaves, and of the characteristics of the soil (i.e. fertilization methods and types) and in relation to the operating conditions of greenhouse cultivation. All this will be achieved through the construction of correlation models between the information previously mentioned and that obtained through classical analysis.