Sensor implementation in horticulture production systems has transcended traditional management practices through the association of multiple environmental factors granting a more optimal use of inputs, resulting in higher crop yield and quality. Complementarily, image-based sensors grant access to plant-specific data to facilitate monitoring plant responses. A fundamental requirement for a decision-making process based on remote sensing is the reliability of the data collected before it is processed into an indirect measurement of plant morpho-physiological response. Extracting meaningful biological insights lies in developing and applying effective processing steps that guarantees features quality. For this reason, our first study implemented and compared three segmentation algorithms: Otsu segmentation, Random Forest classification and Convolutional Neural Network (CNN) U-Net to find the most adaptable segmentation strategy that classifies lettuce canopy pixels more accurately and efficiently under a greenhouse variable illumination condition in a multi-plant hydroponic setup. The results showed that CNN U-Net achieved the best classification performance metrics on validation images at different timepoints and stages of plant development, demonstrating its adaptability. Our second study used an image processing workflow that accurately extracted vertical plant height and canopy geometrical/color features of individual plants from an RGB-D (Red Green Blue Depth) sensor. A simple linear regression and two supervised learning regression models, i.e., LASSO regression and random forest were implemented to estimate ‘Chicarita’ romaine lettuce (Lactuca sativa) leaf fresh weight, leaf dry weight and leaf area grown under greenhouse conditions in a hydroponic system. Prediction accuracy among models for unseen observations was compared were LASSO regression outperformed the other models on the predictive power all three biomass responses as shown by highest R2 and the lowest RMSE. Finally, our imaging methodology was used as a tool to provide morphometric and spectral canopy information extracted from lettuce using color and multispectral imaging as a tipburn monitoring tool. Canopy shape features such as canopy compactness and incident light per canopy extracted allowed us to differentiate morphometric patterns and track growth dynamics for two Romaine lettuce cultivars (Lactuca sativa ‘Chicarita’ and ‘Dragoon’). In a parallel multispectral workflow, band mean reflectance and vegetation index values extracted per plant using a multispectral imaging sensor showed clear patterns across cultivars and light conditions that have the potential to be used as additional criteria for anticipated tipburn spectral characterization.