Scholarlink

This paper presents a computer vision system developed for counting the number of kernels in corn cobs based on several images taken from different angles. A kernel detector was developed that independently analyzes images of cobs. In addition, a linear model was proposed for aggregating results from different images. Special features that do not depend on the order of the object’s photographs reduced the degree of model overtraining. The developed solutions were implemented in SeedMetrics (Brazil) and are used in industrial practice.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automated Corn Kernel Counting from Multiple Views

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mikhail Viktorovich Yurushkin,
  • Artyom Alexandrovich Gornostal,
  • Pavel Vasilyevich Babich,
  • Ricardo Ferreira Domingues,
  • Hugo César Rodrigues Moreira Catão,
  • Reinaldo Scherner

摘要

This paper presents a computer vision system developed for counting the number of kernels in corn cobs based on several images taken from different angles. A kernel detector was developed that independently analyzes images of cobs. In addition, a linear model was proposed for aggregating results from different images. Special features that do not depend on the order of the object’s photographs reduced the degree of model overtraining. The developed solutions were implemented in SeedMetrics (Brazil) and are used in industrial practice.