基于河马优化算法的成行作物导航路径提取方法

打开文本图片集

中图分类号：S233.4 文献标识码：A 文章编号：2095-5553（2025）10-0120-09

Abstract：For green strip-arranged crops such as garlic and green onions，which are planted in rows，a navigation path extraction method based on the Hippopotamus Optimization （HO）algorithm was proposed to address thechalenge of enablingautomatic field movement for harvesting machinery.First，the fieldimages were converted intograyscale， applyingcolor thresholding，and binarizing them using OpenCV.Morphological analysis was then carriedoutbased on the spatial distribution of thecrops.Theoriginal HippoOptimization algorithm wasimprovedusing chaotic mapping techniques andfurtheroptimized basedonthedensitydistributionoffeature points inthe binarized images.Anenhancedversionof the Random Sample Consensus （RANSAC） method was subsequently employed to fit the navigation path from the extracted feature points.Finally，a real-world dataset of fieldcrop images was constructed for evaluation.These images were procesedusing both the improvedandoriginal versionsof the HippopotamusOptimizationalgorithm，aswellasother optimizationmethods for comparativeanalysis.Theresults indicated thatthe improved Hippo Optimizationalgorithm outperformed theoriginal onein optimizing the distribution offeature points within the binarized images.Thenavigation path angle error remained within an absolute value range of 6∘ ，and the average processing time was O.23 s per image. Additioally，the method demonstrated strong adaptabilityacrossvarious green crop types.Thisapproach offeredan efective andautomated navigation path recognition solution for harvesting equipment operating in garlic and onion fields.

Keywords：harvesting machinery；Hippopotamus Optimization algorithm；row crops；image processing；path fiting

0 引言

随着人工智能、物联网大数据等前沿技术的快速发展，农业作业方式正从机械化向自动化、智能化迈进。（剩余14717字）