Visual navigation method for agricultural mobile robots based on spatial continuity clustering algorithm

With the development of the information age, more intelligent products have brought convenience to human life. Robots, as intelligent crystallization products, have been widely used in many fields. Agricultural robots have accelerated the agricultural intelligence. However, visual navigation and trajectory tracking of agricultural robots are still the top priorities of agricultural robot research. The algorithm analysis is carried out for agricultural robot vision navigation trajectory uncertainty. The least squares method is applied to collect robot image data. Then, the linear judgment dimensionality reduction algorithm is used to cluster and analyze the trajectory data of robot visual navigation. According to the experimental results, the linear judgment dimension reduction algorithm outperforms the specific assignment clustering algorithm and the long short-term memory algorithm in robot visual navigation. This error value is 0.73 less than the specific allocation clustering algorithm and 2.69 less than the long short-term memory algorithm. The loss function is less than 0.8 for the specific assignment clustering algorithms, and less than 2.2 for the long short-term memory algorithms. It shows that the whole linear judgment dimension reduction clustering algorithm is obviously superior to the specific assignment clustering algorithm and long short-term memory algorithm. The proposed algorithm has better performance and accuracy, making it more suitable for visual navigation of robots.