| 引用本文: | 张慧,伍萍辉,张馨,等.非垂直拍摄获取细叶作物覆盖度优化算法研究[J].灌溉排水学报,0,():-. |
| ZHANG Hui,WU Pinghui,ZHANG Xin,et al.非垂直拍摄获取细叶作物覆盖度优化算法研究[J].灌溉排水学报,0,():-. |
|
| 摘要: |
| 【目的】探究非垂直拍摄获取细叶作物覆盖度优化算法以及非垂直拍摄对作物覆盖度的影响规律。【方法】本文在北京市小汤山基地草业研究中心露天环境下,以青绿薹草、结缕草为实验对象,获取45°、60°、90°(与地面夹角)三种拍摄角度下的图像,采用基于自适应权重粒子群改进K-means的图像分割方法,分析不同拍摄角度对青绿薹草覆盖度测量精度的影响,研究非垂直拍摄与垂直角度下覆盖度的关系曲线。首先将草RGB图像转化成HSV颜色空间,在HSV颜色空间利用自适应权重PSO算法向全局像素解的子空间搜寻,通过迭代搜寻到全局最优解,确定最佳初始聚类中心；其次利用k均值算法对不同角度图像像素进行聚类划分,从而得到草冠层区域分割结果,最后采用形态学滤波方法对分割结果进行优化。【结果】结果表明：该方法在光照适应性和对抗复杂环境两个方面均优于传统K-means方法,能够克服多角度拍摄导致图像色彩分布不均匀的问题,有效分割出草冠层区域,总体相对误差和均方根误差均在0.2以下,高于传统K-means算法精度。【结论】三个角度下计算出的覆盖度呈显著地线性关系,研究结果可为自然复杂环境下机器视觉多角度监测作物覆盖度提供一种有效途径。 |
| 关键词: 图像分割;K-means算法;PSO优化算法;多角度;覆盖度 |
| DOI： |
| 分类号:S24 |
| 基金项目:国家重点研发计划 2016YFC0403102；北京市农林科学院科技创新能力建设专项(KJCX20170204)；北京市农林科学院科研创新平台建设项目(PT2019-21) |
|
| Research on Optimization Algorithm of Fine Leaf Crop Coverage Based on Non-Vertical Shooting |
|
ZHANG Hui1, WU Pinghui1, ZHANG Xin2, SUN Tiejun3, XUE Xuzhang4, ZHENG Wengang4
|
|
1.School of Electronic Information Engineering,Hebei University of Technology;2.Beijing Research Center for Information Technology in Agriculture;3.Beijing grass industry and environment research center,Beijing academy of agriculture and forestry;4.Beijing Research Center for Information Technology in Agriculture
|
| Abstract: |
| 【Objective】To explore the optimization algorithm of fine leaf crop coverage and the effect of non-vertical shooting on crop coverage in non-vertical shooting.【Method】this paper takes Carex breviculmis R. Br. and Zoysia japonica Steud. as experimental objects in the open environment of the Xiaotangshan Base Grass Research Center in Beijing, and obtains images from three shooting angles of 45°, 60° and 90° (including the Angle with the ground). The image segmentation method based on adaptive weight particle swarm optimization K-means Was used to analyze the influence of different shooting angles on the measurement accuracy of Carex breviculmis R. Br. coverage, and the relationship between non-vertical shooting and coverage under vertical angle is studied. Firstly, the RGB image of grass is transformed into HSV color space. In the HSV color space, the adaptive weight PSO algorithm is used to search the subspace of the global pixel solution. The global optimal solution is found by iterative search to determine the best initial cluster center. Secondly, the k-means algorithm is used to cluster the pixels of different angles to obtain the result of segmentation of the canopy layer. Finally, the segmentation result is optimized by the morphological filtering method.【Results】The results show that the proposed method is superior to the traditional K-means method in both light adaptation and complex environment. It can overcome the problem of uneven color distribution caused by multi-angle shooting and effectively segment the canopy layer area. The relative error and root mean square error are all below 0.2, which is higher than the accuracy of the traditional K-means algorithm.【Conclusion】The calculated coverage under three angles is significantly linear. The research results can provide an effective way for machine vision to monitor crop coverage in a natural complex environment. |
| Key words: image segmentation; K-means algorithm; PSO optimization algorithm; multi-angle; coverage |
