植物叶片图像边缘检测Canny/SUSAN算法比较

Comparison of Image Edge Detection Algorithms Canny/SUSAN for Plant Leaves

植物叶片图像的边缘检测是基于图像分析的植物叶片识别技术中的重要环节.图像边缘检测算法的合理选取与叶片识别的准确度紧密相关.运用Canny算法和SUSAN算法分别对10种具有典型叶形特征的植物叶片图像进行边缘检测,并作主观分析和客观评价.结果表明,植物叶片的叶形和叶脉特征对算法效果的影响较大,在实际应用中应考虑这2种算法对叶形的选择性.

Edge detection for plant leaf image is an important part of image recognition technology,and the selection of the image edge detection algorithm is closely correlated with the accuracy of leaf recognition.In this paper,the Canny algorithm and SUSAN algorithm are employed for plant leaf image edge detection of 10 kinds of plants with typical leaf shapes,and the objective evaluation and subjective analysis are carried out.It is demonstrated that leaf shapes and leaf vein features have significant effect on the effectiveness of the algorithms.Therefore,the selectivity of the two algorithms should be considered in practical applications.