Automated bird detection and targeted activation of ultrasonic bird repeller for wind turbines

Patent metadata

Field	Value
Publication number	US-12408655-B2
Application number	US-202318328802-A
Country	US
Kind code	B2
Filing date	Jun 5, 2023
Priority date	Jun 5, 2023
Publication date	Sep 9, 2025
Grant date	Sep 9, 2025

Abstract

Systems and methods are provided for detecting and repelling bird species approaching a wind farm, using targeted ultrasonic/sonic frequencies. A zone of vulnerability boundary is set around the wind farm at which bird repellers start responding to instructions to broadcast ultrasonic/sonic. Based on the received plurality of images, at least one bird species is identified as approaching the zone of vulnerability boundary. The bird repellers receive instructions to broadcast beginning with the calculated starting ultrasonic/sonic frequency, and incrementally increasing or decreasing the starting ultrasonic/sonic frequency in response to the at least one identified bird species continuing to approach or to move away from the zone of vulnerability boundary. A learning model is updated with statistics, which the identified plurality of bird species and the ultrasonic/sonic frequency receiving the optimal response from the identified plurality of bird species.

First claim

What is claimed is:

1. A method comprising:
receiving a plurality of thermal images from a plurality of thermal imaging cameras, wherein the thermal imaging cameras detect metabolic heat, speed, and direction of approaching birds, and wherein the thermal imaging cameras are placed around and within a perimeter of a wind farm;
based on a geospatial data shapefile describing a topology of the wind farm, setting a configurable zone of vulnerability boundary at which the thermal imaging cameras begin a thermal image capture, and at which bird repellers start responding to instructions to broadcast ultrasonic/sonic;
based on the received plurality of thermal images, identifying by visual recognition at least one bird species approaching the zone of vulnerability boundary;
based on sound sensitivity data corresponding to the identified at least one bird species, instructing the bird repellers to broadcast an ultrasonic/sonic frequency beginning with an optimal frequency value and dynamically and incrementally increasing the ultrasonic/sonic frequency up to a maximum sensitivity limit in response to the at least one identified bird species continuing to approach the zone of vulnerability, or decreasing the ultrasonic/sonic frequency in response to the at least one identified bird species moving away from the zone of vulnerability boundary; and
updating a reinforcement learning model with statistics, the statistics being the identified plurality of bird species and the ultrasonic/sonic frequency receiving the optimal response from the identified plurality of bird species.

2. The method of claim 1 , wherein the zone of vulnerability boundary is set at any distance from a boundary of a wind farm.

3. The method of claim 1 , wherein a plurality of datapoints in a geospatial shapefile identifies a boundary of a wind farm.

4. The method of claim 1 , further comprising calculating a starting ultrasonic/sonic frequency based on the identified at least one bird species, and historical learning model input.

5. The method of claim 1 , wherein the ultrasonic/sonic frequency is not incremented beyond an upper sensitivity limit of the bird species.

6. The method of claim 1 , wherein based on there being more than one bird species, setting a maximum the ultrasonic/sonic frequency to an upper sensitivity limit of the bird species being most sensitive.

7. The method of claim 1 , wherein identifying at least one bird species further comprises:
comparing the received plurality of images to images in a bird knowledge database; and
based on a percent confidence of accuracy, identifying at least one bird species in the bird knowledge database.

8. A computer program product, the computer program product comprising a non-transitory tangible storage device having program code embodied therewith, the program code executable by a processor of a computer to perform a method, the method comprising:
receiving a plurality of thermal images from a plurality of thermal imaging cameras, wherein the thermal imaging cameras detect metabolic heat, speed, and direction of approaching birds, and wherein the thermal imaging cameras are placed around and within a perimeter of a wind farm;
based on a geospatial data shapefile describing a topology of the wind farm, setting a configurable zone of vulnerability boundary at which the thermal imaging cameras begin thermal image capture, and at which bird repellers start responding to instructions to broadcast ultrasonic/sonic;
based on the received plurality of thermal images, identifying by visual recognition at least one bird species approaching the zone of vulnerability boundary;
based on sound sensitivity data corresponding to the identified at least one bird species, instructing the bird repellers to broadcast an ultrasonic/sonic frequency beginning with an optimal frequency value and dynamically and incrementally increasing the ultrasonic/sonic frequency up to a maximum sensitivity limit in response to the at least one identified bird species continuing to approach the zone of vulnerability, or decreasing the ultrasonic/sonic frequency in response to the at least one identified bird species moving away from the zone of vulnerability boundary; and
updating a reinforcement learning model with statistics, the statistics being the identified plurality of bird species and the ultrasonic/sonic frequency receiving the optimal response from the identified plurality of bird species.

9. The computer program product of claim 8 , wherein the zone of vulnerability boundary is set at any distance from a boundary of a wind farm.

10. The computer program product of claim 8 , wherein a plurality of datapoints in a geospatial shapefile identifies a boundary of a wind farm.

11. The computer program product of claim 8 , further comprising calculating a starting ultrasonic/sonic frequency based on the identified at least one bird species, and historical learning model input.

12. The computer program product of claim 8 , wherein the ultrasonic/sonic frequency is not incremented beyond an upper sensitivity limit of the bird species.

13. The computer program product of claim 8 , wherein based on there being more than one bird species, setting a maximum the ultrasonic/sonic frequency to an upper sensitivity limit of the bird species being most sensitive.

14. The computer program product of claim 8 , wherein identifying at least one bird species further comprises:
comparing the received plurality of images to images in a bird knowledge database; and
based on a percent confidence of accuracy, identifying at least one bird species in the bird knowledge database.

15. A computer system, comprising:
one or more processors;
a memory coupled to at least one of the processors;
a set of computer program instructions stored in the memory and executed by at least one of the processors in order to perform actions of:
receiving a plurality of thermal images from a plurality of thermal imaging cameras, wherein the thermal imaging cameras detect metabolic heat, speed, and direction of approaching birds, and wherein the thermal imaging cameras are placed around and within a perimeter of a wind farm;
based on a geospatial data shapefile describing a topology of the wind farm, setting a configurable zone of vulnerability boundary at which the thermal imaging cameras begin thermal image capture, and at which bird repellers start responding to instructions to broadcast ultrasonic/sonic;
based on the received plurality of thermal images, identifying by visual recognition at least one bird species approaching the zone of vulnerability boundary;
based on sound sensitivity data corresponding to the identified at least one bird species, instructing the bird repellers to broadcast an ultrasonic/sonic frequency beginning with an optimal frequency value and dynamically and incrementally increasing the ultrasonic/sonic frequency up to a maximum sensitivity limit in response to the at least one identified bird species continuing to approach the zone of vulnerability, or decreasing the ultrasonic/sonic frequency in response to the at least one identified bird species moving away from the zone of vulnerability boundary; and
updating a reinforcement learning model with statistics, the statistics being the identified plurality of bird species and the ultrasonic/sonic frequency receiving the optimal response from the identified plurality of bird species.

16. The computer system of claim 15 , wherein the zone of vulnerability boundary is set at any distance from a boundary of a wind farm.

17. The computer system of c

Assignees

• IBM

Inventors

• Rakshit Sarbajit K
• Hazra Jagabondhu
• Padmanaban Manikandan

Classifications

• G05B13/0265

  the criterion being a learning criterion · CPC title

• A01M31/002

  Detecting animals in a given area · CPC title

• A01M29/16

  using sound waves · CPC title

• A01M29/10

  using light sources, e.g. lasers or flashing lights · CPC title

• A01M29/18 Primary

  using ultrasonic signals · CPC title