2. From the point of each point (d, d) traversal
(3) from the upper left to lower right traversal (x, y), and meet the d<=x & lt; Width - d d & lt;=y & lt; Height? D, or turn to step 8
4.. Determine the setpoint (x, y) grey value is less than the average grey value of the whole image, if it is, go to step 3
5. Start from this check d values within the scope of the gradient, if the area of the gradient value is greater than alpha, then go to step 3
6. Check whether within the scope of this point d existing feature points, if you have, go to step 3), and then start the next point of traversal
7.. To save the coordinates of points in the feature point set, and add it to the count register, and then go to step 3)
8. After the traversal, judge whether the count value is equal to N, if it is, the end of the algorithm, otherwise, if more than N, then increase the value of d, begin to traverse and judge again, until they are the same, if small, under the condition of the count is greater than N choose the result of the last time, and the application of random refused to count for N feature points, if the first count is less than N, reduce the value of d, this and count is greater than N similar
N is the total number of feature points, the count is to extract the feature points, d is adjusted by the algorithm itself, the radius of the alpha is to distinguish whether point on the edge of the gradient threshold, the width is the width of the image, the height is the height of the image
CodePudding user response:
Recite the text?
CodePudding user response:
That there is no code can retell