Orchard Environmrnt Path Identification Method Research Based On Deep Learning

Fruit trees are planted in rows in orchard environment,which has obvious characteristics of fruit trees rows.And is easy to realize the application of automatic driving between fruit trees rows.The application of intelligent orchard machinery is helpful to reduce manpower input,improve operation accuracy,and improve orchard production efficiency.This study mainly studied the orchard machinery path recognition method based on deep learning.Traditional image processing methods are easily affected by light intensity,environment,and other factors in fruit tree recognition.Deep learning has natural advantages and strong robustness in fruit tree recognition and location.Therefore,it is of great significance to study the path recognition of orchard machinery,which has a practical application prospect.The deep learning method and the least square method combined in this study and proposes the DL-LS(Deep learning-Leas square)algorithm in order to realize the orchard machinery path planning.Many image data collected in actual orchard used for training YOLO V3 after sorting and marking.In orchard environment,there are many interferences such as power poles,fruit tree support poles and weeds in the fruit tree row.At the same time,it is difficult to quickly identify some common tree and nontree objects by using traditional methods.The YOLO V3 network can accurately identify the trunk and fruit trees,and output the coordinates of the bounding box when detecting the image after training.The average detection accuracy of tree trunk and fruit tree is 92.11%.The rectangular frame coordinate information is used to extract the path recognition of orchard machinery walking.The reference point coordinates of this fruit tree can be obtained by sorting and calculating the coordinates of the trunk boundary box.Then,the reference points of the two sides of the fruit trees are fitted with the least square method to get the reference line of the current fruit tree row.Finally,the corresponding points of the two groups of reference lines are taken respectively.And the midpoint of the line segment connected by the two points is calculated.The straight line where the two middle points are located is the orchard machinery route.And it can realize the accurate fitting of the fruit tree line center line.The automatic walking between rows of fruit trees can be realized by getting the mechanical walking route of orchard,which lays a foundation for large-scale orchard management.Distance information from the trunk is also needed to achieve the task of precise spraying pesticide and avoiding resource waste and soil pollution for the spraying orchard machinery.The distance between the orchard machinery and the trunk is measured by taking the midpoint of the rectangular frame of the trunk as the fixed measurement point.