The measurement of phenotypic variables in plants is a fundamental process for growth characterization, assessment of physiological state, and decision-making support in agricultural systems. Traditionally, this process has been labor-intensive, as it relies on manual measurements or high-cost instrumented systems, limiting its applicability in continuous monitoring scenarios. In this context, this work presents a low-cost, low-power, non-destructive methodology based on an embedded ESP32-CAM system for image acquisition and the estimation of phenotypic variables using multiview stereo vision and computer vision techniques. The proposed system is designed to provide an accessible alternative for in-field visual data acquisition and processing without physically altering the plant structure. A multiview stereo vision configuration composed of two ESP32-CAM modules is employed, enabling depth estimation through disparity computation. This approach allows the recovery of spatial information from image pairs. The methodology is based on feature extraction and geometric modeling, establishing relationships between image-based measurements and the actual dimensions of the plants. The proposed approach enables the non-destructive estimation of structural variables, such as height and volume, using low-cost hardware. The accuracy of these measurements is assessed through comparison with ground truth data, validating the system’s suitability for accurate and reliable implementation in plant phenotyping applications.