We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
基于数字图像的棉田复杂背景下棉蚜统计方法.
- Authors
顾佳敏; 王佩玲; 刘阳天; 高攀; 郭文超
- Abstract
【Objective】 This paper aims to present a new automatic counting method for cotton aphids in the hope of achieving the rapid and accurate counting of aphids in cotton fields under complex backgrounds. 【Method】 A large amount of RGB data of cotton aphids was analyzed by a K-means clustering algorithm to obtain an accurate model. Autonomous structural elements were used to complete the corrosion de-noising, and a modulo operation was performed on the number of pixels in the overlapping area. First, the noises were divided into 13 categories according to the colors of images, and the aphids were divided into 7 types. Then, the aphids were classified again after the RGB data of each type was obtained. The data were then analyzed to establish models for the color segmentation of aphids and noises. Next, the association of autonomous structural elements was established according to the principle of statistics, and the optimal structure elements of the images with different noise levels were selected for corrosion de-noising. Finally, the number of cotton aphids in the overlapping area was counted by performing a modulo operation based on the number of pixels of the overlapping area and the expected size of the single-headed aphids. 【Result】 Experimental results showed that the method proposed in this paper can effectively and accurately count cotton aphids in cotton fields under a complex background with an average accuracy of 86.47%. Besides this, in the process of image processing, the dependence of the algorithm on threshold was greatly reduced and the problem of image adhesion segmentation of cotton aphid was solved effectively. Finally, the count of cotton aphid in complex background based on digital image was completed. 【Conclusion】
- Publication
Xinjiang Agricultural Sciences, 2018, Vol 55, Issue 12, p2279
- ISSN
1001-4330
- Publication type
Article
- DOI
10.6048/j.issn.1001-4330.2018.12.015