Automatic mantispid egg detection and counting using image nature

Pei Ying Yang, Chin Dar Tseng, Tai Lin Huang, Chao Hong Liu, I. Hsing Tsai, Wei Chun Lin, Chin Shiuh Shieh, Shyh An Yeh, Stephen Wan Leung, Tsair Fwu Lee*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

Abstract

Mantispids are small brown bugs about 1.5 cm in length. Mantispid eggs are produced in large quantities, with about 1000 eggs per spawning, and are tiny and densely packed. Traditionally, mantispid eggs are counted manually. However, counting such a large quantity of eggs is difficult. To provide accurate data for researchers, we detail methods to accurately detect and count the number of mantispid eggs using image processing. The following methods were used to count the mantispid eggs: background estimation, morphological image processing, background subtraction, stretching, image thresholding, gray-level transformation, labeling and counting. The results of automated counting were compared with the results of manual counting. The segmentation results were verified, and the accuracy of the mantispid egg counts was determined to be 100%. This provides a useful resource for mantispid egg counting. The automatic counting system cannot only count mantispid eggs, it can also be used to count other similar insect eggs.

Original languageEnglish
Title of host publicationPhysics and Mechanics of New Materials and their Application
PublisherWorld Scientific Publishing Co. Pte Ltd
Pages33-38
Number of pages6
ISBN (Electronic)9789819807727
ISBN (Print)9789819807710
DOIs
StatePublished - 01 01 2025

Bibliographical note

Publisher Copyright:
© 2025 by World Scientific Publishing Co. Pte. Ltd.

Keywords

  • automatic counting
  • background estimation
  • counting
  • labeling
  • Mantispid

Fingerprint

Dive into the research topics of 'Automatic mantispid egg detection and counting using image nature'. Together they form a unique fingerprint.

Cite this