基于机器视觉的鸡蛋体积与表面积计算方法

Egg Geometry Calculations Based on Machine Vision

鸡蛋体积与表面积传统手工检测方法速度慢、效率低,不能满足实际生产的需求。利用机器视觉技术,假设理想鸡蛋图像关于纵径对称,提出像素体积和Vp与像素表面积和Sp,给出在数字图像中的计算方法;最后建立了鸡蛋体积V、表面积S与Vp、Sp之间的关系模型。试验验证表明:鸡蛋体积预测模型相关系数为0.965,实测误差±1 cm3内准确率达92%;鸡蛋表面积预测模型相关系数为0.971,实测误差±1 cm2内准确率达88%。

The conventional manual methods for egg volume（V） and surface area＇（S） detection are usually slow,inefficient and not able to meet the actual production requirements,so machine vision technology was researched to replace them.Assumed the ideal egg image was symmetric about longitudinal diameter,the definition of volume pixel（Vp）,surface areas pixel（Sp） and the calculation were proposed.Then,the module between volume（surface area） and Vp（Sp） was established.Experimental results showed that the determination coefficient of volume module and surface area was 0.965 and 0.971,respectively;the test accuracy of egg volume module and surface area reached 92% with ±1cm3 error and 88% with ±1cm2 error respectively.