Backgroud:
A distance-mean-smoothed iterative corn seedling monopoly detection algorithm based on distance-mean-smoothed iteration is proposed to address the problems such as the complexity of traditional crop row line detection for outlier removal algorithms.
Methods
First, based on the green component enhancement method, improved OTSU algorithm and morphological feature analysis noise reduction processing, the image pre-processing such as graying, binarization and morphological noise reduction is performed on the maize seedling graph; then, the binary image with morphological noise reduction is traversed along the horizontal direction row by row to find the starting and ending points of the monopoly, and the DBSCAN algorithm is used for monopoly clustering; finally, the distance-mean-based Finally, the outlier removal algorithm based on the distance-mean iterative algorithm was applied to remove the outliers and linear fitting based on the least squares method was used to obtain the outline of the monopoly.
Results
The experimental results show that the algorithm has a high accuracy in extracting the monopoly furrows in maize seedling stage, and the mean value of the fitted linear determination coefficient is 0.98s.
Conclusions
This paper provides a new method for detecting monopoly furrows in maize seedlings in the field; The image results processed by this algorithm have a good fit and can reflect the general orientation of the rows in the field.