Tomato fruit maturity segmentation method and picking robot

Abstract

The invention provides a tomato fruit maturity segmentation method and a picking robot, and the method comprises the steps: improving a region-based mask convolutional neural network algorithm trunk network, optimizing a mask branch loss function, and obtaining an improved region-based mask convolutional neural network model; reducing the possibility of repetition in the feature information integration process; and training the improved region-based mask convolutional neural network model according to different types of tomato picture training sets to obtain a recognition model, obtaining a fruit picture of a tomato fruit, recognizing the fruit picture through the recognition model to obtain a segmentation result of the fruit picture. The tomato fruit identification efficiency and accuracyare improved.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a tomato fruit ripeness segmentation method and a picking robot. Background technique [0002] With the continuous adjustment of my country's agricultural structure and the rapid development of the Internet, facility fruit and vegetable planting has gradually shifted from traditional greenhouses to intelligent greenhouses, and the output has also increased day by day. Fruit and vegetable picking has attracted people's attention, especially tomato. At present, tomato picking in facility agriculture mainly relies on manual work. Due to the large greenhouse area and the huge number of tomatoes picked, it will lead to a lot of waste of manpower and material resources, high labor intensity and low work efficiency. Tomato picking robots can reduce operators and production costs, and have practical significance for the automation of picking operations in facility agricul...

AI Technical Summary

Problems solved by technology

[0004] The invention provides a tomato fruit maturity segmentation method, a picking robot and a computer-readable storage medium, which are used to solve the defect that the fruit in different growth stages cannot be instance-segmented in the prior art, and realize the segmentation of fruits in different growth stages. Instance segmentation of tomato fruit to improve the accuracy of tomato fruit recognition and segmentation

Images