孢子
算法
粘附
聚类分析
数学
索贝尔算子
噪音(视频)
人工智能
计算机科学
图像处理
生物
材料科学
图像(数学)
植物
边缘检测
复合材料
作者
W. H. Tao,Tao Cheng,Dongyan Zhang,Gensheng Hu,Gan Zhang,Chen Gu,Yanbing Xue,Xin‐Gen Zhou
标识
DOI:10.1109/agro-geoinformatics59224.2023.10233682
摘要
Aiming at the problem of low efficiency of manual counting of wheat scab spore microscopic images, we propose a counting algorithm for adhesion of wheat scabs spore based on contour angle-to-distance ratio. Specifically, the algorithm first improves the microscopic image quality of wheat scab spores utilizing Mean Shift and eliminates the influence of impurities and noise. Then, the adhesion of spores were screened by the Sobel operator and shape feature factor, and all adhesion spores were segmented according to the angle-to-distance ratio of spore contour, and finally, he spore count was completed. The experimental results show that this algorithm achieves an average accuracy of 93.1% in 313 wheat scab spore microscopic images, which is 7.9% higher than the traditional machine learning clustering algorithm. This algorithm can quickly and accurately calculate the number of wheat scab spores and can provide technical support for disease prevention and food security.
科研通智能强力驱动
Strongly Powered by AbleSci AI