Flower thinning for fruit trees in time is an important operation to keep a suitable quantity of fruits and guarantee the quality of fruits. Accurate detection of flower density is the premise of precise flower thinning, and machine vision provides an effective approach to achieving the accurate identification of flower density. To detect the flower density on the proximal side of Y-shaped densely planted peach trees accurately, this study proposed a method based on an RGBD camera and a convolutional neural network that incorporated an attention mechanism and multi-scale feature fusion. Firstly, image acquisition and preprocessing were performed with the RGBD camera, and the complex background and distal flowers were filtered out through depth information. Then, a convolutional neural network for flower density detection based on an attention mechanism and multi-scale feature fusion, named the flower counting network (FC-Net), was constructed and tested. Results showed that the coefficient of determination (R2) between the estimated number of flowers by the FC-Net and the real values reached 0.95, the mean absolute error (MAE) was 4.3, the root mean square error (RMSE) was 5.65, the counting error rate (Er) was 0.02%, and the processing time of one image was 0.12 s. The proposed FC-Net can provide visual support for intelligent mechanical flower thinning operations.
When the orchard operation platform is in use within the orchard, issues of tilting and overturning can arise due to uneven ground, necessitating instant leveling. In this study, the orchard operation platform is simplified into a four-point leveling mechanism, and an adaptive leveling system based on an inertial measurement unit (IMU) is designed. The relationship between coordinate transformation is utilized to derive the platform tilt angle and the position error relationship of the electric actuator, allowing for the analysis of the angle adjustment factors of the leveling mechanism. Through co-simulation using MATLAB and ADAMS, fuzzy control is implemented in addition to PID control, resulting in improved performance. A prototype model of the orchard operation platform is produced and tested, with the platform’s attitude angle remaining stable within a range of ±1.5°. The average leveling time is found to be within 3.6 s. The mean values of dynamic leveling inclination under PID and fuzzy PID control are 2.6° and 1.6°, respectively, with corresponding standard deviations of 1.4° and 0.8°. It conforms to the development trend of agricultural machinery electrification and intelligence and provides a reference basis for manufacturers.