传感器翻译Major colour method. Given the transformed nhin Eq. , t

传感器翻译
Major colour method. Given the transformed nh
in Eq. , the recognition to rapeseed seeds colour can be
achieved by the following methods:
For single seed, we firstly calculated the histogram of nh
of each pixel of a seed in an image. The maximum of the
histogram was the major colour.
For an image or a sample, we also calculated the histogram
of major colour of each seed. And the maximum of
the histogram was the major colour. The ratio of the
maximum of the histogram and the total pixels of a seed
was major colour rate.
In order to compare the performances of two colour
recognition methods, all seeds images of 206 sample
images were tested. The colour recognition results of two
methods were summarized in Fig.3. The recognition
result of major colour method showed that there were five
classes colour of seeds, i.e., black, red, orange, yellow
and blue. And the recognition result of common colour
method showed that there were eight classes colour of
seeds, i.e., black, red, orange, yellow, green, cyan, blue
and magenta. But in fact, there should be four classes
colour of seeds, i.e., black, red, orange and yellow. The
colour recognition accuracy of common method and
major colour method were 83.96% and 98.91%, respectively.

主要色彩方法.在均衡器中转换n个色调,通过下列方法可以得出油菜籽的种源色：
对于单颗种子,我们首先计算出在某幅图像中某个种子的每一个像素的n个色调直方图,直方图中的最大值就是主要色彩.
对于某个图像或样品,我们也可以计算出每颗种子主要色彩的直方图.直方图的最大值就是主要色彩.直方图最大值的比率和种子整个像素的比率就是主要色彩的比率.
为了比较两种色彩识别方法的性能,206个样品图像的所有种子图像都进行测试.两种方法的色彩识别结果归纳为图3.主要色彩方法的识别结果表明,有五个级别的种子,即黑色、红色、橙色、黄色、蓝色.常见的色彩识别方法的识别结果表明,有八种级别的种子色彩,黑色、红色、橙色、绿色、青色、蓝色、洋红.但实际上,应该有四个级别的种子色彩,即黑色、红色、橙色、黄色.常用方法和主要色彩方法的色彩识别精确度分别是83.96%和98.91%.