Method and system of recognizing object edges and computer-readable storage medium
A method and system of recognizing object edges and a computer-readable storage medium are provided. The method includes: obtaining an input image; recognizing the input image and obtaining a relative position of each object vertex and a corresponding image vertex thereof; determining a reference position of each object vertex in the input image according to the relative position of each object vertex and the corresponding image vertex thereof; performing corner point detection in a predetermined area where the reference position of the object vertex is located for each object vertex; determining an actual position of each object vertex in the input image according to a result of the corner point detection; and sequentially connecting adjacent object vertices to form edge lines to obtain the edges of the object with edges in the input image according to the actual position of each object vertex in the input image.
1. A method of recognizing object edges, comprising:
obtaining an input image, wherein the input image comprises an object with edges, and the edges of the object comprise a plurality of object vertices, wherein the plurality of object vertices comprise a first object vertex and a second object vertex;
recognizing, through an object vertex recognition model, the input image and obtaining a first relative position between the first object vertex and a first image vertex, and obtaining a second relative position between the second object vertex and a second image vertex;
determining a reference position of each of the object vertices in the input image according to the relative position of each of the object vertices and the corresponding image vertex thereof;
performing a corner point detection in a predetermined area where the reference position of the object vertex is located for each of the object vertices;
determining an actual position of each of the object vertices in the input image according to a result of the corner point detection; and
sequentially connecting adjacent object vertices to form edge lines to obtain the edges of the object with edges in the input image according to the actual position of each of the object vertices in the input image.
2. The method of recognizing object edges according to claim 1 , wherein the predetermined area where the reference position of the object vertex is located is a circular area with a pixel point at the reference position of the object vertex as a center of the circle and a first predetermined pixel as a radius, and
the step of performing the corner point detection in the predetermined area where the reference position of the object vertex is located for each of the object vertices further comprises:
performing the corner point detection on the pixel point in the circular area corresponding to each of the object vertices; treating all of the pixel points whose feature value variation ranges are greater than a predetermined threshold as candidate corner points during a process of the corner point detection; and determining a target corner point corresponding to each of the object vertices among the candidate corner points.
3. The method of recognizing object edges according to claim 2 , wherein the step of determining the target corner point corresponding to each of the object vertices among the candidate corner points further comprises:
step A 1 , sorting the candidate corner points in a descending order according to the feature value variation ranges, wherein the candidate corner point ranked first is determined as the target corner point, and the candidate corner point ranked second is determined as a current corner point to be selected;
step A 2 , determining whether all distances between the current corner point to be selected and all current target corner points are greater than a second predetermined pixel, wherein step A 3 is performed if all distances between the current corner point to be selected and all current target corner points are greater than the second predetermined pixel is determined, otherwise, step A 4 is performed;
step A 3 , determining the current corner point to be selected as the target corner point; and
step A 4 , discarding the current corner point to be selected, determining the candidate corner point ranked next as the current corner point to be selected, and performing step A 2 again.
4. The method of recognizing object edges according to claim 3 , wherein a range of the second predetermined pixel is ≥50 pixels.
5. The method of recognizing object edges according to claim 2 , wherein a range of the first predetermined pixel is 10 to 20 pixels.
6. The method of recognizing object edges according to claim 2 , wherein the step of determining the actual position of each of the object vertices in the input image according to the result of the corner point detection further comprises:
for each of the object vertices, if the result of the corner point detection of the object vertex comprises a corner point, determining a position of the corner point as the actual position of the object vertex in the input image; or if the result of the corner point detection of the object vertex does not comprise a corner point, determining the reference position of the object vertex in the input image as the actual position of the object vertex in the input image.
7. The method of recognizing object edges according to claim 1 , wherein the step of obtaining the first relative position between the first object vertex and a first image vertex, and obtaining a second relative position between the second object vertex and a second image vertex further comprises:
obtaining the first relative position between the first object vertex and the first image vertex closest to the first object vertex in the input image.
8. The method of recognizing object edges according to claim 7 , wherein the step of determining the reference position of each of the object vertices in the input image according to the relative position of each of the object vertices and the corresponding image vertex thereof further comprises:
converting the first relative position of the first object vertex and the first image vertex closest to the first object vertex in the input image into coordinates of the first object vertex in a target coordinate system to obtain the reference position of the first object vertex in the input image for each of the object vertices.
9. The method of recognizing object edges according to claim 8 , wherein an origin of the target coordinate system is a position point of the input image.
10. The method of recognizing object edges according to claim 1 , further comprising:
correcting the object with edges by using the edge lines.
11. The method of recognizing object edges according to claim 10 , wherein four object vertices are provided and four edge lines are provided, which are a first line, a second line, a third line, and a fourth line in sequence, the first line and the third line are arranged in a first direction, the second line and the fourth line are arranged in a second direction, and the first direction is perpendicular to the second direction,
the step of correcting the object with edges by using the edge lines further comprises:
obtaining projection lengths of the first line and the third line in the second direction separately and obtaining projection lengths of the second line and the fourth line in the first direction separately;
calculating an average value of the projection lengths of the first line and the third line in the second direction to act as a corrected length of the object with edges in the second direction; and
calculating an average value of the projection lengths of the second line and the fourth line in the first direction to act as a corrected length of the object with edges in the first direction.
12. The method of recognizing object edges according to claim 1 , wherein the object vertex recognition model is a neural network model.
13. A system of recognizing object edges, comprising a processor and a memory, wherein the memory stores a command, and the steps of the method according to claim 1 are implemented when the command is executed by the processor.
14. A non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores a command, and the steps of the method according to claim 1 are implemented when the command is executed.