Fire detection method and apparatus

1. A fire detection method comprising:

acquiring RGB data from an image;

converting the RGB data into HSL data;

correcting the converted HSL data;

extracting a fire color region from the corrected HSL data;

obtaining the indexes using the RGB data and the corrected HSL data; and

determining whether a fire has occurred using the indexes,

wherein when all of H, S and L of the HSL data are HIGH, a corresponding pixel is a fire color region and thus is set to HIGH; and when at least one among H, S and L is LOW, a corresponding pixel is set to LOW, and

wherein the correcting of the HSL data comprises changing a corresponding pixel from LOW to HIGH when only one among H, S and L is LOW and pixels surrounding the pixel are HIGH.

2. The method of claim 1 , wherein the indexes comprises a periodicity index representing a periodicity of the fire color region, a moving object index representing a movement region of the fire color region, and a zero cross index representing a variation of flicker in the fire color region.

3. The method of claim 2 , wherein the periodicity index is generated in such a manner as to obtain a first operation value through a mask operation with respect to a current frame belonging to a block, to obtain second operation values through a mask operation with respect to the other frames belonging to the block, to obtain differences between the first operation value and the second operation values, and to obtain a ratio of a standard variance of the differences and an average of the differences.

4. The method of claim 2 , wherein the moving object index is generated by accumulating a moving object pixel count of all frames in a corresponding block in such a manner as to obtain an RGB average with respect to the respective pixels of a current frame, to obtain, according to each pixel, a variance corresponding to a difference between the obtained RGB average and an RGB average of a corresponding pixel of a previous frame, and to increase the moving object pixel count by one when the variance exceeds a preset threshold value.

5. The method of claim 2 , wherein the zero cross index is generated by accumulating a zero cross count of all frames in a corresponding block in such a manner as: to generate a flicker variation value through gray values of a pervious-previous frame, a pervious frame, and a current frame with respect to a specific pixel; and to increase the zero cross count by one when the flicker variation value exceeds a lower limit value or an upper limit value in a next frame at a state in which the flicker variation value has exceeded the upper limit value or the lower limit value, respectively, in a current frame.

6. The method of claim 1 , wherein the determining of whether a fire has occurred comprises:

comparing a periodicity index with a corresponding threshold value;

comparing a moving object index with a corresponding threshold value;

comparing a zero cross index with a corresponding threshold value;

determining that a fire is suspected in a corresponding block when it is determined in all the comparing that a fire is suspected; and

determining that a fire has occurred when blocks determined to be suspected as a fire exist successively, and the number of the successive blocks is equal to or greater than a preset fire reference value.

7. The method of claim 6 , wherein the determining of whether a fire has occurred further comprises:

checking a distance between the moving object and a corresponding saturation; and

checking a distance between the moving object and a previous moving object.

8. A fire detection method comprising:

partitioning an image into blocks which comprises a plurality of frames;

acquiring RGB data from the image;

converting the RGB data into HSL data;

correcting the converted HSL data;

extracting a fire color region from the corrected HSL data;

obtaining one or more indexes related to attributes of a fire using the RGB data and the corrected HSL data;

determining whether a fire is suspected using the indexes with respect to the blocks; and

determining that a fire has occurred when blocks determined to be suspected as a fire exist successively, and the number of the blocks is equal to or greater than a fire reference value,

wherein when all of H, S and L of the HSL data are HIGH, a corresponding pixel is a fire color region and thus is set to HIGH; and when at least one among H, S and L is LOW, a corresponding pixel is set to LOW,

wherein the correcting of the HSL data comprises changing a corresponding pixel from LOW to HIGH when only one among H, S and L is LOW and pixels surrounding the pixel are HIGH.

9. The method of claim 8 , wherein the indexes comprises a periodicity index representing a periodicity of the fire color region, a moving object index representing a movement region of the fire color region, and a zero cross index representing a variation of flicker in the fire color region.

10. The method of claim 9 , wherein the periodicity index is generated in such a manner as to obtain a first operation value through a mask operation with respect to a current frame belonging to a block, to obtain second operation values through a mask operation with respect to the other frames belonging to the block, to obtain differences between the first operation value and the second operation values, and to obtain a ratio of a standard variance of the differences and an average of the differences.

11. The method of claim 9 , wherein the moving object index is generated by accumulating a moving object pixel count of all frames in a corresponding block in such a manner as to obtain an RGB average with respect to the respective pixels of a current frame, to obtain, according to each pixel, a variance corresponding to a difference between the obtained RGB average and an RGB average of a corresponding pixel of a previous frame, and to increase the moving object pixel count by one when the variance exceeds a preset threshold value.

12. The method of claim 9 , wherein the zero cross index is generated by accumulating a zero cross count of all frames in a corresponding block in such a manner as: to generate a flicker variation value through gray values of a pervious-previous frame, a pervious frame, and a current frame with respect to a specific pixel; and to increase the zero cross count by one when the flicker variation value exceeds a lower limit value or an upper limit value in a next frame at a state in which the flicker variation value has exceeded the upper limit value or the lower limit value, respectively, in a current frame.

13. A fire detection apparatus comprising:

an index unit for generating one or more indexes related to attributes of a fire using RGB data which is acquired from an image;

a fire determination unit for determining whether a fire has occurred using the indexes, which have been generated by the index unit; and

an HSL unit for converting the RGB data into HSL data and correcting the converted HSL data;

wherein, among the corrected HSL data, a pixel corresponding to a fire color region is set to HIGH, and the other pixels are set to LOW,

wherein the indexes comprises a periodicity index representing a periodicity of the fire color region, a moving object index representing a movement region of the fire color region, and a zero cross index representing a variation of flicker in the fire color region,

wherein the periodicity index is generated in such a manner as to obtain a first operation value through a mask operation with respect to a current frame belonging to a block, to obtain second operation values through a mask operation with respect to the other frames belonging to the block, to obtain differences between the first operation value and the second operation values, and to obtain a ratio of a standard variance of the differences and an average of the differences.

14. The apparatus of claim 13 , wherein the moving object index is generated by accumulating a moving object pixel count of all frames in a corresponding block in such a manner as to obtain an RGB average with respect to the respective pixels of a current frame, to obtain, according to each pixel, a variance corresponding to a difference between the obtained RGB average and an RGB average of a corresponding pixel of a previous frame, and to increase the moving object pixel count by one when the variance exceeds a preset threshold value.

15. The apparatus of claim 13 , wherein the zero cross index is generated by accumulating a zero cross count of all frames in a corresponding block in such a manner as: to generate a flicker variation value through gray values of a pervious-previous frame, a pervious frame, and a current frame with respect to a specific pixel; and to increase the zero cross count by one when the flicker variation value exceeds a lower limit value or an upper limit value in a next frame at a state in which the flicker variation value has exceeded the upper limit value or the lower limit value, respectively, in a current frame.

16. The apparatus of claim 13 , wherein the fire determination unit determines whether a fire is suspected according to each block of the image using the indexes; and determines that a fire has occurred when blocks determined to be suspected as a fire exist successively, and the number of the blocks is equal to or greater than a preset fire reference value.