Description: 我们给出一个模板 和一幅图象 。不难发现原图中左边暗，右边亮，中间存在着一条明显的边界。进行模板操作后的结果如下： 。
可以看出，第3、4列比其他列的灰度值高很多，人眼观察时，就能发现一条很明显的亮边，其它区域都很暗，这样就起到了边沿检测的作用。
为什么会这样呢？仔细看看那个模板就明白了，它的意思是将右邻点的灰度值减左邻点的灰度值作为该点的灰度值。在灰度相近的区域内，这么做的结果使得该点的灰度值接近于0；而在边界附近，灰度值有明显的跳变，这么做的结果使得该点的灰度值很大，这样就出现了上面的结果。
这种模板就是一种边沿检测器，它在数学上的涵义是一种基于梯度的滤波器，又称边沿算子，你没有必要知道梯度的确切涵义，只要有这个概念就可以了。梯度是有方向的，和边沿的方向总是正交(垂直)的，例如，对于上面那幅图象的转置图象，边是水平方向的，我们可以用梯度是垂直方向的模板 检测它的边沿。
例如，一个梯度为45度方向模板 ，可以检测出135度方向的边沿。-we give a template and an image. It is not difficult to find the maximum were left dark, right-liang, in the middle there is a clear boundary. After the template for the operation results are as follows :. Can be seen, three, four out other than the gray value is much higher, eye observation, we can obviously found a bright side. Other regions are dark, and this has played a role in the detection of 2500. Why is this the case? A closer look at the template on which to understand it. It means the right to the point o gray minus left point as a gray value of the point of gray values. In a similar gray area, do so as a result of the point of gray values close to 0; And near the border. gray values jump significantly changed, the results do make the point very gray value, and this appeared to Platform: |
Size: 9333 |
Author:李涯 |
Hits:
Description: 我们给出一个模板 和一幅图象 。不难发现原图中左边暗，右边亮，中间存在着一条明显的边界。进行模板操作后的结果如下： 。
可以看出，第3、4列比其他列的灰度值高很多，人眼观察时，就能发现一条很明显的亮边，其它区域都很暗，这样就起到了边沿检测的作用。
为什么会这样呢？仔细看看那个模板就明白了，它的意思是将右邻点的灰度值减左邻点的灰度值作为该点的灰度值。在灰度相近的区域内，这么做的结果使得该点的灰度值接近于0；而在边界附近，灰度值有明显的跳变，这么做的结果使得该点的灰度值很大，这样就出现了上面的结果。
这种模板就是一种边沿检测器，它在数学上的涵义是一种基于梯度的滤波器，又称边沿算子，你没有必要知道梯度的确切涵义，只要有这个概念就可以了。梯度是有方向的，和边沿的方向总是正交(垂直)的，例如，对于上面那幅图象的转置图象，边是水平方向的，我们可以用梯度是垂直方向的模板 检测它的边沿。
例如，一个梯度为45度方向模板 ，可以检测出135度方向的边沿。-we give a template and an image. It is not difficult to find the maximum were left dark, right-liang, in the middle there is a clear boundary. After the template for the operation results are as follows :. Can be seen, three, four out other than the gray value is much higher, eye observation, we can obviously found a bright side. Other regions are dark, and this has played a role in the detection of 2500. Why is this the case? A closer look at the template on which to understand it. It means the right to the point o gray minus left point as a gray value of the point of gray values. In a similar gray area, do so as a result of the point of gray values close to 0; And near the border. gray values jump significantly changed, the results do make the point very gray value, and this appeared to Platform: |
Size: 9216 |
Author:李涯 |
Hits:
Description: 人眼的识别是计算机人脸识别和智能监控中的重要部分。本文提出了一种基于灰度投影
曲线和边缘检测进行人眼定位的新算法。首先利用原图的垂直灰度投影曲线确定人脸的左右边界, 之
后利用人脸区域的水平灰度投影确定头顶至鼻中部的上下边界, 然后利用人脸比例模型估计人眼的大
概位置, 分析人眼区域的边缘图象, 给出人眼的确切位置。实现表明, 该算法简单、快速, 鲁棒性强-Identification of the human eye is the computer face recognition and intelligent control of the important part. This paper presents a projection based on gray curve and edge detection to locate the new algorithm for the human eye. Firstly, original gray projection curve to determine the vertical face of the left and right boundaries, and then use the face region to determine the level of gray projection head to the nose and down the middle of the border, then use the model to estimate the proportion of people face around the eye position analysis the edge of the eye region image, given the exact location of the human eye. Realize that the algorithm is simple, fast and robust Platform: |
Size: 11167744 |
Author:廖减员 |
Hits:
Search - eye border detection