Design of Vehicle Image Stabilization System and Imaging Quality Impact Test for Typical Grasslands

Grassland intelligence perception is conducive to adequate monitoring of grassland and saving human resources. However, grassland road conditions are complex, and vibration often disturbs image acquisition sensors. This study explores the influencing factors of green vegetation imaging. According to the analysis of variance, when the collection equipment is far away from the vegetation, shooting from an overhead angle, and the vegetation density is low, the obtained data has a higher proportion of precise data and a smaller proportion of blurred and distorted images. This study proposed a theoretical model of the impact of vibration on green vegetation imaging. It was found that the overall imaging quality was better when the vibration displacement was within 50 μm, and the vibration frequency was within 15 Hz. This study constructed a complete vehicle image stabilization system. This study proposed a new PSO-based GRC neural network tuning fuzzy PID control method. This study conducted tests based on vibration data obtained by driving the chassis on the grassland. In short grass areas, the image stabilization system performed best, with the maximum vibration displacement during compensation being around 13μm. For vibrations in other grassland road conditions, the vibration displacement after compensation by this system was within 50μm, meeting the above requirements for stable imaging. The image stabilization system proposed in this study can select mechanism parameters and optimization algorithms according to grassland conditions. This can provide a reference for image stabilization systems under different grassland conditions. This study may contribute to grassland intelligence perception robot.