Vibration Shapes Identification Applying Eulerian Video Magnification on Coffee Fruits to Study the Selective Harvesting

Abstract

In the last few years, Colombian coffee production has decreased by different economic factors related to the increment of costs in manual selective harvest operations. Consequently, an interest has arisen for studying selective harvesting methods based on mechanical vibrations in seeking for an alternative technology. In this sense, it is mainly essential to understand the dynamic behavior of the coffee fruits at different ripening stages. In this study, an experiment was proposed to observe the vibration shapes of coffee fruits (Coffea arabica var. Castillo) in defined frequency intervals. The ripening stages were categorized by a strategy based on color measurements in the CIELab space. Four ripening stages were determined from color data taken of 84 fruits, which allowed the development of the dynamic experiments over classified fruits. The dynamic excitation was applied to the coffee fruits between 0–1000 Hz in which ripe fruits evidenced higher vibration amplitudes in specific frequency intervals. These motions cannot be visualized from a normal video-camera (60 Fps). Therefore, a high-speed camera (Chronos 1.4) was used to record these during harmonic excitation. Eulerian video magnification technique was applied to the video-based measurements to visualize the amplified vibration shapes. The magnified motions were contrasted with vibrometry measurements and a piezo sensor attached to the fruit pedicel. Results show that the ripe fruit can vibrate with higher amplitudes in comparison with the other ripening stages, this information could be used to harvest selectively only ripe fruits.