Navigation Line Extraction Method for Broad-Leaved Plants in the Multi-Period Environments of the High-Ridge Cultivation Mode

Navigation line extraction is critical for precision agriculture and automatic navigation. A novel method extracting navigation lines based on machine vision proposed herein using a straight detected high-ridge crop row. Aiming at the low-level automation of machines in field environments cultivation mode broad-leaved plants, suitable multiple periods with high timeliness designed. The comprises four sequentially linked phases: image segmentation, feature point extraction, calculation, dynamic segmentation horizontal strip number feedback. a* component CIE-Lab colour space extracted to preliminarily extract row features. OTSU algorithm combined morphological processing completely separate rows backgrounds. points are an improved isometric segmented vertical projection method. While calculating lines, adaptive clustering used cluster adjacent points. determine final sets, sets optimised lateral distance methods. In optimisation process, linear regression Huber loss function fit set obtain centreline, calculated according two lines. Finally, before entering next frame feedback mechanism calculate strips introduced improve ability adapt periods. experimental results show that can meet efficiency requirements visual average time samples 38.53 ms. Compared least squares method, longer growth period crops.