Image processing apparatus and method

1. An image processing apparatus comprising at least one processor to implement:

an information obtainer configured to obtain an image created by photographing a region of interest and location information indicating a location from which the image is photographed;

a density map generator configured to generate a crowd density map of the image based on crowd density information estimated from the image;

a perspective map generator configured to generate a perspective map that provides distance information indicating a real-life distance between two points in the image based on the location information; and

a heat map generator configured to generate a heat map that displays the crowd density map on a map of the region of interest based on the distance information,

wherein the density map generator is further configured to:

split the image into patch images respectively representing different parts of the region of interest, the different parts being of a same real-life size,

estimate a crowd density class of each of the patch images based on characteristic information extracted from each of the patch images, and

estimate the crowd density information based on the crowd density class.

2. The image processing apparatus of claim 1 , wherein the density map generator is configured to resize the patch images to have a same pixel size, and estimate the crowd density class of the patch images using the resized patch images.

3. The image processing apparatus of claim 1 , wherein the density map generator is configured to calculate a probability with respect to each of previously set crowd density classes using the characteristic information of the patch images and classify each of the patch images based on the probability.

4. The image processing apparatus of claim 1 , wherein the density map generator is configured to allocate different visual properties to the patch images based on the crowd density class.

5. The image processing apparatus of claim 1 , wherein the perspective map shows multiple horizontal lines, and

wherein a real-life distance between any two adjacent horizontal lines from among the multiple horizontal lines is uniform.

6. The image processing apparatus of claim 1 , wherein the heat map generator is configured to adjust transparency of the crowd density map that overlaps the map of the region of interest.

7. The image processing apparatus of claim 1 , wherein the heat map generator is configured to merge crowd density information of a plurality of crowd density maps at an overlapping region according to a preset merge condition and display the merged crowd density information on the map of the region of interest.

8. The image processing apparatus of claim 1 , wherein the location information comprises GPS information and altitude information.

9. The image processing apparatus of claim 1 , wherein the location information indicates a location of an aerial vehicle in which a camera that created the image is mounted.

10. An image processing method performed by an image processing apparatus, the image processing method comprising:

obtaining an image created by photographing a region of interest and location information indicating a location from which the image is photographed;

generating a crowd density map of the image based on crowd density information estimated from the image;

generating a perspective map that provides distance information indicating a real-life distance between two points in the image based on the location information; and

generating a heat map that displays the crowd density map on a map of the region of interest based on the distance information,

wherein the generating of the crowd density map comprises:

splitting the image into patch images respectively representing different parts of the region of interest, the different parts being of a same real-life size;

estimating a crowd density class of each of the patch images based on characteristic information extracted from each of the patch images; and

estimating the crowd density information based on the crowd density class.

11. The image processing method of claim 10 , wherein the estimating comprises:

resizing the patch images to have a same pixel size; and

estimating the crowd density class of the patch images using the resized patch images.

12. The image processing method of claim 10 , wherein the estimating of the crowd density class comprises:

calculating a probability with respect to each of previously set crowd density classes using the characteristic information of the patch images and classifying the patch images based on the probability.

13. The image processing method of claim 10 , wherein the generating of the crowd density map comprises:

allocating different visual properties to the patch images based on the crowd density class.

14. The image processing method of claim 10 , wherein the perspective map shows multiple horizontal lines, and

wherein a real-life distance between any two adjacent horizontal lines from among the multiple horizontal lines is uniform.

15. The image processing method of claim 10 , wherein the generating of the heat map comprises:

adjusting transparency of the crowd density map that overlaps the map of the region of interest.

16. The image processing method of claim 10 , wherein the generating of the heat map comprises:

merging crowd density information of a plurality of crowd density maps at an overlapping region according to a preset merge condition and displaying the merged crowd density information on the map of the region of interest.

17. The image processing method of claim 10 , wherein the location information comprises GPS information and altitude information.

18. A non-transitory computer-readable recording medium having recorded thereon a computer program for executing the image processing method of claim 10 .