Validation and Training of AI-Algorithms for Precision Greenhouse Agriculture

The client has an application to assist in the algorithm-based decision-making process. From here, they can use this service to test the robustness of the algrithms and their ability to gneralize in decision-making. This will allow to: (i) test the algorithm's functionality and accuracy of measurement and decision-making on a crop for the first time, (ii) test the algorithm's functionality and accuracy assessment under different conditions, (iii) test the algorithm's functionality and accuracy of measurement and decision-making on new crops.

As a general rule, the service will be linked to the complete production cycle of the plant species under trial. For exmple, currently, greenhouse horticultural crops are sown between April and May, and harvested in September. A complete cycle can last between 5 and 6 months. The service is provided at the following location: https://maps.app.goo.gl/hBMw4W7vGsA1sxkm6. There is no requirement for the client's location. The client will receive an agronomic report on the crop's progress and the data for training and/or validating the algorithm. In order to provide this data, the client must provide the necessary equipment and technical support for its installation. The equipment will depend on the parameters of interest for automatic data collection or for establishing potential correlations between them. By mutual agreement with the service requester, the data will be provided in the most appropriate standard format, either raw or processed, especially in the case of treatments with replicates, to ensure the robustness of the information provided. Intellectual property and data rights will be reflected in a document signed by both parties, such that intellectual property will belong to the entity that generates it, data will be transferred under non-exclusive use agreements, and the pre-existing know-how of each entity will be preserved.

The trials can be customised based in the characteristics of interest for the study, the types of sensors, the crop species and varieties, etc. Therefore, the applicant must agree with the service provider n the terms and scope of the studies. The maximum number of trials per client will also be agreed upon, taking into accunt the number f species of interest, the diversity f measurement paramentes, and the duration of the trials. For example, the Horticultural greenhouse production roughly follows the following growing cycle: from April-May to September. Therefore, this period will be appropriate for data collection. It is highly recommended to visit the facilities before providing the service.