计算机视觉鸭蛋重量检测方法研究

为了实现鸭蛋重量的智能检测,提出基于计算机视觉鸭蛋重量检测方法。首先构造鸭蛋图像的灰度—梯度共生矩阵,根据最大熵原理求出最佳灰度和梯度分割阈值,实现二维阈值分割;其次采用数学形态学方法进行分割图像后处理,去除蛋壳表面存在的伪目标,统计鸭蛋区域像素点;最后利用多项式拟合方法建立面积与鸭蛋重量关系。试验表明,检测误差在±2g以内,平均误差为-0.13353 g,检测精度能够满足生产加工需求。

Study on Duck's Egg Weight Detection Methods Based on Computer Vision

In order to achieve the intelligent detection of duck's egg weight,a method to detect duck's egg weight based on computer vision was proposed.First,the gray level-gradient co-occurrence matrix of the duck's egg image was formulated.Then,the optimum gray level and gradient segmentation thresholding was obtained according to maximum entropy principle,thus two-dimensional thresholding segmentation was realized;second,segmentation images were post-processed by using the method of mathematical morphology,so the phony targets on eggshell were removed,pixel points in duck's egg region were counted;finally,the relationship between areas and duck's egg weight was set up by using polynomial fitting method,the experiment showed that the detection error was within ± 2 g,average error was-0.133 53 g,and the detection accuracy can met the needs of production and processing.