Plant phenotypes can be obtained from dedicated phenotyping platforms, from farmer’s fields, or even from ecological diagnostics in natural environment. Plant phenotypes encompass all the traits that can be measured  like morphological traits, phenological, quality, sensibility to biotic & abiotic stress, yield component traits and their linked eco-systemic services. Trait measurements are done at different scales, from the field to the cell. Phenotyping platforms measure a wide range of structural and functional plant traits at the same time as collecting meticulous metadata on the environment and experimental setup [Fiorani and Schurr, 2013]. 

Big data are produced in the form of alphanumeric matrix, images, readings which may be complemented with knowledge such as statistical or 3D models.  Like all agronomic data, the phenotyping data: 

1. to the germplasm information
2. have to be associated with environmental observations
   
3. are often produced in the frame of well defined experimental designs which include crop management practices 
   
4. need intensive multi-scale integration (field/plot/plant, cell/tissue/plant, etc.) iv. are often assessed along plant development periods. 
   

It is vital that pre-breeders, breeders, agronomists, crop modellers, climate-change scientists, share a common language to describe phenotypes and also interpret descriptions provided by farmers for their preferred varieties' performance.

However, the lack of shared frameworks for the acquisition, storage and data management of phenotypic data currently hampers optimized comparisons of crop management scenarios and limits the efficiency of  breeding.  A community effort is thus necessary for achieving and validating a full set of standards and best practices for annotations that will support scientists and data managers.
To improve this situation, several bioinformatics projects have developed standards, tools and data models to address the issues encountered by scientists and to promote the harmonization and sharing of data; but a complete solution is not yet fully developed, not are the existing solutions universally applied.
Data interoperability enables comparison and interpretation of crop trait data across evaluation sites or phenotyping platforms but is impeded by semantic heterogeneity of ways variables are named.

An extended use of ontologies, metadata and existing standards is proposed. Reference Plant Ontologies (e.g. Plant Ontology, Phenotypic Quality Ontology) are being developed in the context of the Planteome project, and are being integrated with the Crop Ontology, which provides harmonized breeders’ trait names, measurement methods, scales and standard variables for currently 20 crops: cassava, banana, barley, chickpea, common bean, cowpea, groundnut, lentil, maize, oat, pearl millet, pigeon pea, potato, rice, sorghum, soybean, sweet potato, wheat and yam. Work is also underway for defining the minimal information for capturing the metadata needed to describe an experiment (e.g. the Minimal Information About a Plant Phenotyping Experiment specification), and data formats (e.g. ISA-tab) for implementing these standards in a standard way to promote data exchange.