Back to results
US 11,109,586 B2Granted
System and methods for automated wildlife detection, monitoring and control
AssigneeBird Control Group BV
InventorsSteinar Henskes, Pim Tammes, Kris Mooi
Life of this patent
Estimating the term…
Abstract
The present disclosure describes a system which is able to detect and recognize wildlife, and in particular birds, using camera images. The present solution is comprised of algorithms, software and integrated hardware devices. Properly equipped, the system can be made to be portable and can be set up at any location for different wildlife detection and repelling purposes.
Claims20
A bird deterrence system comprising:
- a camera device configured to acquire a sequence of images of a region of interest;
- a first central processing device including a Convoluted Neural Network (CNN) which employs an unsupervised learning method to create a background mask containing backgrounds made up of pixels within said sequence of images that have a low probability of changing over time said processing device detecting pixel values that differ significantly from said background mask to detect birds within said images of said region of interest, said first central processing device creating a data set relative to detected birds within said images of said region of interest;
- a second central processing device running a software application which is programmed to:
- receive said data set from said first central processing device;
- and
- generate a repelling action effective for targeting detected birds within said region of interest based on said data set;
- a device configured to produce a repelling action;
- a control unit configured to control the device,
- wherein the control unit is programmed to:
- receive said generated repelling action;
- and
- control the device consistent with said generated repelling action.
9 more elements
A bird deterrence system comprising:
- a camera device configured to acquire a sequence of images of a region of interest;
- a first central processing device including a Convoluted Neural Network (CNN) which employs an unsupervised learning method to create a background mask containing backgrounds made up of pixels within said sequence of images that have a low probability of changing over time said processing device detecting pixel values that differ significantly from said background mask detect birds within said images of said region of interest, said first central processing device creating a data set relative to said detected birds within said images of said region of interest;
- a second central processing device running a software application which is programmed to:
- receive said data set from said first central processing device;
- and
- creating a repelling pattern effective for targeting detected birds within said region of interest based on said data set;
- a device configured to produce a light beam;
- a control unit configured to control the device,
- wherein the control unit is programmed to:
- receive said repelling pattern;
- control movement of the device along a trajectory consistent with said repelling pattern;
- and
- activate the device to produce a moving light beam along the trajectory.
10 more elements
A method of detecting birds in an image comprising the steps of:
- acquiring with a camera, a sequence of images of a region of interest;
- providing said sequence of images to a central processing unit that is integrated with said camera, said central processing unit running a software application that is operative for
- receiving said sequence of images,
- separating the images into a foreground and a background using an unsupervised learning method to create a background mask containing backgrounds made up of pixels within said sequence of images that have a low probability of changing over time;
- morphologically processing said images to decrease noise;
- and
- detecting objects within said images using a contour detection algorithm which identifies pixel values that differ significantly from said background mask.
4 more elements
A method of detecting birds in an image comprising the steps of:
- acquiring with a camera, a first sequence of images of a region of interest;
- providing said first sequence of images to a central processing unit that is integrated with said camera;
- ground truth labelling objects in each of said images in said first sequence of images;
- said central processing unit running a Convolutional Neural Network (CNN) with an object detection architecture;
- training said CNN using an unsupervised learning method to create a background mask containing backgrounds made up of pixels within said first sequence of images that have a low probability of changing over time;
- acquiring with said camera a second sequence of images of said region of interest;
- providing said second sequence of images to said central processing unit said central processing unit;
- receiving said second sequence of images,
- detecting objects within said second images using said Convolutional Neural Network wherein a pixel that differs over time is identified as belonging to a foreground mask, and
- classifying said objects that are detected as said foreground mask.
7 more elements
Dependencies are inferred from the claim text.