Research On The Method Of Judging Litchi Maturity And Size Based On Image Deep Learning

Object detection is the documentation of an object in the image laterally with its localization and classification.It has inclusive range applications and critical component for visionbased software systems.Automatic fruit picking is a challenging problem in robotics with wide application field.A requirement for realization of a robotic fruit picker is its ability to detect fruits in tree tops.An expert system,which would be able to compete with human perception,must be skillful of recognizing fruits among leaves and branches under uncontrolled conditions,where fruits are occluded and shaded.On the other hand,the localization accuracy can easily be affected by unstructured growing environments with variable illumination conditions and unpredictable growing shapes of weight-bearing stems carrying litchi clusters.Locating picking points plays a vital role in robotic litchi harvesting in orchards.Litchi Fruit is detected in images using a state-of-the-art,Faster R-CNN detector.We adopt this model,through finetuning,for the task of litchi fruit based detection.The main objective of current research is to design an efficient and accurate Faster-RCNN based detection of litchi fruits.The test results show the proposed algorithm has achieved higher detection accuracy and lower processing time than the traditional detectors.Current work demonstrated that we introduce a new,high-quality dataset of images containing litchi fruit.This work presents a new multi-sensor framework to identify resourcefully,track,localize,and map every piece of fruit in a commercial Litchi orchard.The performances of the several investigated techniques were assessed based on accuracy and efficiency.The proposed faster R-CNN method proves to be efficient and also produce highly accurate and consistent results.