【摘要】 掌握牛肉肌肉的颜色特征有助于进一步的图像分析。为此,利用C#平台编制程序,对牛肉图像中的肌肉像素进行了随机抽样,并用统计学中的矩估计法和极大似然法对肌肉的HSL各分量的上下限进行了界定,最后进行图像分割来证实这些数据的正确性。结果证实,用亮度值(L)或饱和度值(S)来区分图像中的肌肉和脂肪更合适。更多还原

【Abstract】 To grasp the HSL character of beef pixel will contribute to image analysis further. Using the software designed with C#, this paper first conducts a random sampling of beef pixel, then defines the upper and lower limit value of hue, saturation, luminance by applying the Moment Estimate and Maximum Likelihood Estimate methods, and finally, demonstrates the accuracy of these data with image segmentation. The result shows that luminance or saturation value of the pixel is a better choice for distinguishing fat from muscle.更多还原