Fig. 2

Pictures of the Earbox system. A The Earbox acquisition system, which allows the simultaneous acquisition of six maize ears on six sides via the rotation of the ear by motorized rollers. The identification of individual ears or ear lots is done by a keyboard or a barcode scanner. The system consists of aluminum profiles, compact laminates and assembly parts supplied by Elcom SAS or manufactured by Phymea Systems (CNC machining or 3D printing). The acquisition system is composed of two Raspberry Pi, each driving a Pi NoIR (V2.0) camera module, and a custom-made Arduino like board (ATMEGA 328P), to control the lighting and the two stepper motors (door and rollers) via two A4988 drivers. The master Rapsberry Pi (model 3 B+) hosts the main Python program, which centralizes all the functions of the system: the graphical user interface via the PyGame library, the control of the slave Rapsberry Pi (model B+) via SSH protocol, the communication with the Arduino like board for motor control, and the saving of the pictures to an external hard drive. B The Earbox system imaging cabin. Polarized lenses are added to the Pi Noir cameras. The lighting system consists of flexible LED strips in the visible (CRI 90) and infrared (940nm) wavelengths behind a frosted polycarbonate diffuser. Rubber strips are added to the rollers for optimal adhesion between the rollers and the ears.