ENBIS-26 Conference

Multivariate SPC of Smart Greenhouses for Berry Production: A Hybrid SPC–ML Framework — Moroccan Case Study

Not scheduled

20m

Centro Didattico Morgagni

Statistical Process Monitoring

Speaker

Abdellah AMZIL (Computer, Networks, Mobility and Modeling Laboratory (IR2M), Faculty of Sciences and Techniques, Hassan First University of Settat)

Description

Berry greenhouses in the Souss-Massa region of Morocco sustain high-value exports of strawberries, blueberries, and raspberries, but their yield and fruit quality are highly sensitive to microclimate deviations. Dense IoT sensor networks generate high-dimensional, autocorrelated, and non-stationary data that violate classical Shewhart, CUSUM, and MEWMA assumptions.

We propose a hybrid Statistical Process Monitoring framework combining a regularized autoencoder, trained on in-control periods, with an adaptive MEWMA chart on reconstruction residuals. The approach is validated on several months of real greenhouse data (temperature, humidity, CO₂, soil moisture, PAR, fertigation) and benchmarked against Hotelling's T², PCA-MSPC, and isolation forests. Results show faster drift detection, better out-of-control ARL, and clearer interpretability through residual contribution plots. We discuss explainable AI and standardization for SPM in precision agriculture.