Data cleaning for an intelligent greenhouse

The effectiveness of greenhouse control can be improved by the application of model based intelligent control. However for this a good model of a greenhouse is needed. For a large variety of industrial or recreational greenhouses the derivation of a fully blown analytical model is not feasible and simplified models serve no practical purpose. Thus black-box modeling has to be applied. Identification (learning) of black-box models requires large amount of data from real greenhouse environments. After recording long time series of greenhouse measurements to serve its purpose the data has to be checked for validity. Measurement errors or missing values are common and must be eliminated to use the collected data efficiently as training samples for the greenhouse model. This paper discusses problems of cleaning the measurement data collected in a well instrumented greenhouse, and introduces solutions for various kinds of missing data problems.