Method and apparatus for detecting smoke from image
View Patent ↗Provided are a fire detecting apparatus and a method thereof for detecting a fire, the method includes operations of extracting a feature of at least one object in an input image by using a value of a brightness difference between pixels of the input image or by using an RGB value of the input image; converting the extracted feature of the at least one object into an N dimensional feature; and performing Support Vector Machine (SVM) machine learning on the N dimensional feature of the at least one object.
1. A fire detecting apparatus comprising:
at least one processor to implement:
an object feature extracting unit configured to extract a feature of at least one object in an input image;
a converting unit configured to convert the extracted feature of the at least one object into an N dimensional feature, where N is a natural number; and
a mechanical learning unit configured to perform machine learning on the N dimensional feature of the at least one object, based on coordinate values of the N dimensional feature, to detect a fire and or a smoke,
wherein the object feature extracting unit is further configured to obtain a differential image of the input image, calculate a weight and a differential direction for each pixel of the differential image, generate a normalization image by normalizing a histogram of the differential direction with respect to the weight, and extract the feature of the at least one object from the normalized histogram.
2. The fire detecting apparatus of claim 1 , wherein the object feature extracting unit extracts the feature of the at least one object by using a value of a brightness difference between pixels of the input image or by using an RGB value of the input image.
3. The fire detecting apparatus of claim 2 , wherein the object feature extracting unit calculates the value of the brightness difference between the pixels of the input image, by using a Histogram of Oriented Gradients (HoG).
4. The fire detecting apparatus of claim 1 , wherein, when the input image contains a smoke image, the object feature extracting unit extracts the feature of the at least one object by using a Histogram of Oriented Gradients (HoG).
5. The fire detecting apparatus of claim 1 , wherein, when the input image contains a fire image, the object feature extracting unit extracts the feature of the at least one object by using a Histogram of Oriented Gradients (HoG) and an RGB value of the input image.
6. The fire detecting apparatus of claim 1 , wherein the converting unit is configured to convert the extracted feature of the at least one object into the N dimensional feature by using a Kernel Mapping method, and
wherein the machine learning is a Support Vector Machine learning.
7. The fire detecting apparatus of claim 1 , wherein the input image is an image obtained by a mobile device while the mobile device is moving.
8. A method of detecting a fire, the method performed by a fire detecting apparatus and comprising:
extracting a feature of at least one object in an input image;
converting the extracted feature of the at least one object into an N dimensional feature, where N is a natural number; and
performing machine learning on the N dimensional feature of the at least one object, based on coordinate values of the N dimensional feature, to detect a fire and/or a smoke,
wherein the extracting comprises:
obtaining a differential image of the input image;
calculating a weight and a differential direction for each pixel of the differential image; and
generating a normalization image by normalizing a histogram of the differential direction with respect to the weight, and extract the feature of the at least one object from the normalized histogram.
9. The method of claim 8 , wherein the extracting comprises extracting the feature of the at least one object by using the value of the brightness difference between the pixels of the input image or by using the RGB value of the input image.
10. The method of claim 9 , wherein the extracting comprises calculating the value of the brightness difference between the pixels of the input image, by using a Histogram of Oriented Gradients (HoG).
11. The method of claim 8 , wherein, when the input image contains a smoke image, the extracting comprises extracting the feature of the at least one object by using a Histogram of Oriented Gradients (HoG).
12. The method of claim 8 , wherein, when the input image contains a fire image, the extracting comprises extracting the feature of the at least one object by using a Histogram of Oriented Gradients (HoG) and an RGB value of the input image.
13. The method of claim 8 , wherein the converting the extracted feature of the at least one object into the N dimensional feature is performed by using a Kernel Mapping method, and
wherein the machine learning is a Support Vector Machine learning.
14. The method of claim 8 , wherein the input image is an image obtained by a mobile device while the mobile device is moving.