Smart IP camera with color night mode

1. An apparatus comprising:

a camera configured to capture color images of a monitored field of view in response to visible light during daytime operation and monochrome infrared images of the monitored field of view in response to infrared light during nighttime operation; and

a processor circuit configured to extract color information of one or more color features of one or more of objects and backgrounds from the color images of said monitored field of view captured during daytime operation and add color from the color features to corresponding monochrome features of one or more corresponding objects and backgrounds detected in the monochrome infrared images of said monitored field of view captured during nighttime operation, wherein said processor circuit (a) stores and executes a first artificial neural network to extract said color information about said one or more color features of said one or more objects and backgrounds from said color images of said monitored field of view captured during a daytime operating mode and stores the color information about the one or more color features of said one or more objects and backgrounds from the color images in a database during the daytime operating mode, (b) stores and executes a second artificial neural network to extract one or more extracted features of one or more of objects and backgrounds from said monochrome infrared images of said monitored field of view captured during a nighttime operating mode, (c) selects one or more color features in the database associated with one or more of the objects and backgrounds from the color images that correspond to the one or more extracted features of one or more of the objects and backgrounds from the monochrome infrared images during the nighttime operating mode, and (d) stores and executes a third artificial neural network to colorize each monochrome infrared image to a full color image by adding the color information associated with the color features selected from the database to the one or more extracted features of the one or more of the objects and backgrounds of the monochrome infrared images corresponding to the one or more of the objects and backgrounds from the color images associated with the color features.

2. The apparatus according to claim 1 , wherein said camera comprises an RGB image sensor and an IR image sensor.

3. The apparatus according to claim 1 , wherein said camera comprises an RGB-IR image sensor.

4. The apparatus according to claim 1 , wherein said color information about said one or more color features extracted from the color images stored in the database comprises pixel values for one or more of an entire image, a region of interest, an object, a background region, or a predetermined portion of an image.

5. The apparatus according to claim 1 , wherein said processor circuit is further configured to dynamically adjust one or more stored color features stored in the database to adapt to a position of the camera changing.

6. The apparatus according to claim 1 , wherein said processor circuit is further configured to dynamically adjust one or more objects stored in the database to adapt to an object in the monitored field of view of the camera changing.

7. The apparatus according to claim 1 , wherein the database is stored locally in a memory connected to said processor circuit.

8. The apparatus according to claim 1 , wherein the database is stored remotely in a cloud resource.

9. The apparatus according to claim 1 , wherein said camera and said processor circuit are part of at least one of a surveillance system, a security system, and an access control system.

10. A method for implementing a color night mode in a smart IP camera comprising:

capturing color images of a monitored field of view in response to visible light during daytime operation using a camera;

capturing monochrome infrared images of the monitored field of view in response to infrared light during nighttime operation using the camera;

extracting color information of one or more color features of one or more of objects and backgrounds from the color images of said monitored field of view captured during the daytime operation using a processor circuit to store and execute a first artificial neural network to extract the color information about the one or more color features of the one or more of objects and backgrounds from said color images and store the color information about the one or more color features in a database during a daytime operating mode;

extracting one or more monochrome features of one or more of objects and backgrounds from said monochrome infrared images of the monitored field of view captured during the nighttime operation using the processor circuit to store and execute a second artificial neural network to extract one or more extracted monochrome features of one or more of objects and backgrounds from said monochrome infrared images and select one or more color features in the database associated with one or more of the objects and backgrounds from the color images that correspond to the one or more extracted monochrome features of the objects and backgrounds from the monochrome infrared images during a nighttime operating mode; and

adding color from the color features extracted during the daytime operation to corresponding monochrome features detected in the monochrome infrared images of the monitored field of view captured during the nighttime operation using the processor circuit to store and execute a third artificial neural network to colorize each monochrome infrared image back to a full color image by adding the color information associated with the color features selected from the database to the one or more extracted monochrome features of the one or more of the objects and backgrounds of the monochrome infrared images corresponding to the one or more of the objects and backgrounds from the color images associated with the color features selected from the database.

11. The method according to claim 10 , wherein:

the color information about the one or more color features extracted from the color images stored in the database comprises pixel values for one or more of an entire image, a region of interest, an object, a background region, or a predetermined portion of an image.

12. The method according to claim 11 , further comprising:

dynamically adjusting one or more color features stored in the database to adapt to a mounting position of the camera changing.

13. The method according to claim 10 , further comprising:

capturing the color images using an RGB image sensor; and

capturing the monochrome infrared images using an IR image sensor.

14. The method according to claim 10 , further comprising:

capturing the color images and the monochrome infrared images using an RGB-IR image sensor.

15. The method according to claim 11 , wherein extracting color features from the color images comprises using the processor circuit to apply a segmentation neural network model to the captured color images.

16. The method according to claim 11 , wherein extracting color features from the color images further comprises using the processor circuit to apply an object detection neural network model to the captured color images to select one or more regions of interest.

17. The method according to claim 11 , wherein adding color from the color features stored in the database to corresponding monochrome features detected in the monochrome infrared images comprises using the processor circuit to apply a fully convolutional neural network model to the captured monochrome infrared images.

18. The method according to claim 17 , wherein the processor circuit applies a fully convolutional neural network model trained with monochrome images and color images to the captured monochrome infrared images.

19. The apparatus according to claim 1 , wherein said first artificial neural network comprises a first segmentation neural network, said second artificial neural network comprises a second segmentation neural network, and said third artificial neural network comprises a fully convolutional neural network.

20. The apparatus according to claim 19 , wherein the third artificial neural network comprises a fully convolutional neural network model trained with monochrome images and color images.