Intelligent micro flight sensing system for detecting the internal and external quality of apples on the tree

Quality detection of apples on trees is important for predicting apple maturity, harvest time and crop yield. Based on the multispectral sensor and red green blue-depth (RGB-D) camera carried by micro unmanned aerial vehicles (UAVs), this study designed a close detection system for the determination of the internal and external quality of apples. An RGB-D camera was used to perceive the spatial position of the apple and predict the apple diameter. The multispectral sensor was self-optimized, inexpensive, small and easy to install in the acquisition module for apple soluble solid content (SSC) detection. The coefficient of the correlation between apple size values obtained by automatic and manual measurements of the system is 0.9672. The root mean square error was 2.2624 mm. The apple SSC detection test includes spectrum acquisition experiments under static and dynamic UAV condi-tions. The results showed that the linear model established by partial least squares regression method performed slightly better than the nonlinear model established by support vector regression method under both conditions. The dynamic modeling results were slightly worse than the static modeling results, but there was little difference between the two sets of modeling results. The correlation coefficient of the prediction set of the model was 0.8568 and the root mean square error was 0.7753 degrees Brix. Results revealed that the flight sensor system designed in this study has the potential to detect the internal and external quality of apples on trees.

Automated summary

A close detection system was designed using a multispectral sensor and an RGB-D camera carried by micro unmanned aerial vehicles (UAVs) for determining the internal and external quality of apples on trees. The system utilized the RGB-D camera to perceive the spatial position and predict the diameter of the apple, and the self-optimized multispectral sensor for apple soluble solid content (SSC) detection. The results showed that the system has the potential to accurately detect the quality of apples on trees.