CN115299428A - Intelligent bird system that drives of thing networking based on degree of depth study - Google Patents
Intelligent bird system that drives of thing networking based on degree of depth study Download PDFInfo
- Publication number
- CN115299428A CN115299428A CN202210932069.2A CN202210932069A CN115299428A CN 115299428 A CN115299428 A CN 115299428A CN 202210932069 A CN202210932069 A CN 202210932069A CN 115299428 A CN115299428 A CN 115299428A
- Authority
- CN
- China
- Prior art keywords
- bird
- module
- bird repelling
- deep learning
- internet
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000006855 networking Effects 0.000 title claims description 17
- 230000001846 repelling effect Effects 0.000 claims abstract description 95
- 238000001514 detection method Methods 0.000 claims abstract description 45
- 238000013135 deep learning Methods 0.000 claims abstract description 41
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 14
- 238000012544 monitoring process Methods 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 13
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 12
- 230000009545 invasion Effects 0.000 claims abstract description 11
- 238000001914 filtration Methods 0.000 claims abstract description 9
- 238000005457 optimization Methods 0.000 claims abstract description 6
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims abstract description 5
- 241000271566 Aves Species 0.000 claims description 40
- 230000005540 biological transmission Effects 0.000 claims description 21
- 238000011176 pooling Methods 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 13
- 238000005286 illumination Methods 0.000 claims description 12
- 238000000034 method Methods 0.000 claims description 12
- 238000004891 communication Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000007635 classification algorithm Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 5
- 239000008186 active pharmaceutical agent Substances 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 210000004205 output neuron Anatomy 0.000 claims description 3
- 241000607479 Yersinia pestis Species 0.000 claims description 2
- 239000003990 capacitor Substances 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 241000272534 Struthio camelus Species 0.000 claims 1
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 241000894007 species Species 0.000 description 6
- 238000007781 pre-processing Methods 0.000 description 5
- 239000012212 insulator Substances 0.000 description 4
- 229910052742 iron Inorganic materials 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000033001 locomotion Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- UXOWGYHJODZGMF-QORCZRPOSA-N Aliskiren Chemical compound COCCCOC1=CC(C[C@@H](C[C@H](N)[C@@H](O)C[C@@H](C(C)C)C(=O)NCC(C)(C)C(N)=O)C(C)C)=CC=C1OC UXOWGYHJODZGMF-QORCZRPOSA-N 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 229960004601 aliskiren Drugs 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 239000011324 bead Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 235000005770 birds nest Nutrition 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 210000003746 feather Anatomy 0.000 description 1
- 230000002431 foraging effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 230000036544 posture Effects 0.000 description 1
- 230000002940 repellent Effects 0.000 description 1
- 239000005871 repellent Substances 0.000 description 1
- 230000031068 symbiosis, encompassing mutualism through parasitism Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 235000005765 wild carrot Nutrition 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01M—CATCHING, TRAPPING OR SCARING OF ANIMALS; APPARATUS FOR THE DESTRUCTION OF NOXIOUS ANIMALS OR NOXIOUS PLANTS
- A01M29/00—Scaring or repelling devices, e.g. bird-scaring apparatus
- A01M29/06—Scaring or repelling devices, e.g. bird-scaring apparatus using visual means, e.g. scarecrows, moving elements, specific shapes, patterns or the like
- A01M29/10—Scaring or repelling devices, e.g. bird-scaring apparatus using visual means, e.g. scarecrows, moving elements, specific shapes, patterns or the like using light sources, e.g. lasers or flashing lights
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01M—CATCHING, TRAPPING OR SCARING OF ANIMALS; APPARATUS FOR THE DESTRUCTION OF NOXIOUS ANIMALS OR NOXIOUS PLANTS
- A01M29/00—Scaring or repelling devices, e.g. bird-scaring apparatus
- A01M29/16—Scaring or repelling devices, e.g. bird-scaring apparatus using sound waves
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01M—CATCHING, TRAPPING OR SCARING OF ANIMALS; APPARATUS FOR THE DESTRUCTION OF NOXIOUS ANIMALS OR NOXIOUS PLANTS
- A01M29/00—Scaring or repelling devices, e.g. bird-scaring apparatus
- A01M29/16—Scaring or repelling devices, e.g. bird-scaring apparatus using sound waves
- A01M29/18—Scaring or repelling devices, e.g. bird-scaring apparatus using sound waves using ultrasonic signals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/36—Applying a local operator, i.e. means to operate on image points situated in the vicinity of a given point; Non-linear local filtering operations, e.g. median filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Birds (AREA)
- Zoology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Environmental Sciences (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Insects & Arthropods (AREA)
- Pest Control & Pesticides (AREA)
- Wood Science & Technology (AREA)
- Human Computer Interaction (AREA)
- Optics & Photonics (AREA)
- Nonlinear Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an intelligent bird repelling system based on the internet of things for deep learning, which takes STM32F as a core, a Doppler radar is combined with a camera, when a moving object is detected by the Doppler radar, information is transmitted to the STM32F, the camera is started to shoot a picture, gaussian filtering and binarization processing are carried out on the picture through an OpenCV computer vision library, the accuracy of target classification detection is improved, a designed convolutional neural network model is utilized, an Adam optimization algorithm is adopted to carry out deep learning, whether the invading object is a bird or not is judged, if a bird repelling mode is started for the bird, the bird repelling is carried out by adopting sound waves and green light flashing in the daytime, the bird repelling is carried out by adopting green light flashing and ultrasonic waves at night, the bird repelling mode is more humanized, remote monitoring of the internet of things is carried out, and a terminal user monitors on a PC or an APP in real time. The device that drives the bird when no birds invasion is dormant mode, has reduced the electric quantity loss, and solar panel and battery combine, are showing and have improved duration, have realized that no external power supply independently drives the bird for a long time, and simple and practical drives bird accurately, has reduced fortune dimension degree of difficulty and cost.
Description
Technical Field
The invention relates to the technical field of Internet of things, in particular to an intelligent bird repelling system based on deep learning for the Internet of things.
Background
Nowadays, with the improvement of global electrification and informatization level, the stability requirement of industry, especially manufacturing industry, on an electric power system is higher and higher, and in order to ensure the safe and stable operation of the electric power system, the occurrence of transmission line faults and control accidents must be reduced, and the safety of the transmission line is related to national safety, economic development and social stability. In recent years, with the increasing number of birds, the range of motion is expanded, some birds put bird nests on poles (towers) of power transmission lines to cause short circuits, various birds like to fall on cross arms of central lines of iron towers after foraging, a large amount of excrement is discharged when the birds stop eating the birds, insulators are polluted, the insulating strength of the insulators is reduced to cause line tripping, the birds often drill into gaps of transformers or alternating current filters to cause instantaneous faults of the power transmission lines, flashover tripping of the power transmission lines is on the trend of rising year by year, power transmission line accidents caused by the birds not only cause huge economic losses, but also seriously disturb normal social production and living orders, and bird damage becomes one of 3 important reasons for the faults of the power transmission lines.
At present, the main bird repelling measures in China are that bird repelling thorns, windmills, insulator umbrella skirts and the like are arranged on poles (towers) of power transmission lines. The bird thorn and windmill devices can not actively detect birds, only the passive blind bird repelling mode is single, a certain bird repelling effect can be achieved in a short time, but the bird repelling mode is long and adaptive to the birds, so that the bird repelling mode becomes a place where the birds inhabit and even nest the birds. In addition, a small amount of voice bird repelling and ultrasonic bird repelling devices are available in domestic markets, but the detection methods are easily influenced by raindrops, fallen leaves, weather, temperature and other environments, the false alarm rate and the missing alarm rate are extremely high, accurate bird repelling cannot be achieved, people are easily disturbed, and environmental pollution is caused, so that a long-term independent working efficient bird repelling device is urgently needed for the power transmission line.
In order to solve the problems, for example, a Chinese patent with publication number CN113383763A discloses an all-weather all-area intelligent bird repelling system and device, which are composed of four parts, namely a detection sensing module, an information comprehensive processing module, a multi-energy bird repelling module and a human-computer interaction module. The detection sensing module consists of a bird detection radar, a photoelectric system, a servo control system and a multifunctional rotary table; the comprehensive processing module adopts a high-speed processing controller hardware architecture to complete information acquisition and processing, data fusion, artificial intelligent model calculation, intelligent control and deep learning; the multifunctional driving-away module adopts a strong-sound bird-repelling method and a blue-light bird-repelling method, and different driving-away multiple combination modes are continuously executed; the man-machine interaction module can intelligently monitor bird target information and send instructions according to the information; the system and the device have the comprehensive precaution functions of intelligently detecting, sensing and acquiring data, fusing and processing the data, performing artificial intelligence edge calculation, performing various driving modes of sound and light, and finally performing continuous iterative optimization and the like according to driving effect evaluation and machine deep learning.
If a Chinese patent with publication number CN114242080A discloses a transformer substation distributed bird repelling method and terminal based on bird voiceprint characteristics, a bird database is established according to bird information collected in advance, and a deep learning network is trained based on the bird database, so that the deep learning network can separate and identify the bird voiceprint information, wherein the bird information comprises the bird voiceprint information and bird species; receiving voiceprint information and bird position information sent by a front-end detection device, separating and identifying bird voiceprint information in the voiceprint information through a deep learning network, and obtaining corresponding bird types; inquiring a bird database according to the bird species, and generating a driving-away strategy corresponding to the bird species according to an inquiry result; sending a driving-away strategy to one distributed bird repeller closest to the position in a plurality of distributed bird repellers distributed at different positions according to the bird position information, and driving the birds by the distributed bird repellers according to the driving-away strategy; the bird repelling device is more targeted, and the bird repelling effect is improved while the bird repelling effectiveness is guaranteed.
At present, the existing bird repelling technology has the following defects: the first patent obtains data through intelligent detection perception, carries out artificial intelligence edge calculation after fusing data, opens the bird equipment that drives that corresponds the position when necessary and drives the bird, carries out sound, CD-ROM drive mode to reach the purpose of automatic bird that drives, regard flying bird driving time as the object degree of depth study, improve birds recognition rate and drive the bird effect. The second patent establishes a bird database by collecting bird information, separates and identifies different bird species through a deep learning network, and selects a corresponding driving-away strategy according to the different species and bird positions. The bird repelling effect is improved to a certain extent by the two patents, however, the bird probing Lei Dayi used by the first patent is mistakenly reported under the influence of environmental factors such as wind, rain and the like, the second patent is separated and identifies different bird types through a deep learning network, a lot of computer computing power is consumed, the accuracy of bird information identification is not high, a battery is used as a power supply, the endurance time cannot be guaranteed, the operation and maintenance difficulty is high, and the prior art still needs to be improved.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides an intelligent bird repelling system based on the Internet of things for deep learning, and aims to solve the problems.
In order to achieve the purpose, the invention is realized by the following technical scheme:
the utility model provides a bird system is driven to thing networking intelligence based on deep learning, by STM32F microcontroller, detection module, illumination detection module, drive bird module, power module and thing networking remote monitoring and patrol and examine module, communication module, survey and drive bird scope and enlarge the module and constitute, the detection module includes Doppler radar, camera, STM32F microcontroller is the whole core module who drives bird system, with detection module, illumination detection module, thing networking remote monitoring and patrol and examine module, power module, communication module, survey and drive bird scope and enlarge the module and be connected, the BC95 module is chooseed for use to communication module, the OV7725 module is chooseed for use to the camera.
Furthermore, the Doppler radar is combined with the OV7725 camera module, when the Doppler radar detects a moving object, detection information is transmitted to the STM32F microcontroller, the OV7725 camera is started to shoot pictures, whether the invading object is birds or not is judged based on a deep learning classification algorithm, and if the invading object is birds, a bird repelling mode is started.
Further, the deep learning target classification algorithm identifies whether an invasive object is a bird or not through a convolutional neural network model, 2 convolutional layers, 2 pooling layers and 2 full-link layers are used, 6 convolutional neural networks are totally adopted, the number of output neurons is 2, gaussian filtering and binarization processing are carried out on a background image and a positive sample image through an OpenCV (open circuit vehicle vision library), the background is filtered, the position and contour features of the invasive object are extracted, graph segmentation is carried out, then a network model is introduced for training and testing, and the size formula of the output feature graph of the first convolutional layer is as follows:
(equation 1):
o-size of output image;
i-size of input image;
k-the size of the convolution kernel;
p is the padding number;
s-moving step length.
Further, the output of the convolution neural network model first layer convolution layer is used as the input of the second layer pooling layer, the output of the second layer pooling layer is obtained according to the filter size and the step length of the second layer pooling layer, the output characteristic image size is obtained according to the number of convolution kernels, the convolution kernel size and the step length in the third layer convolution layer, the output obtained through the last layer pooling layer is connected with the full-connection layer to spread tensor data into vector data, and the vector data is connected with the full-connection layer through the full-connection layer and the ftsomax activation function is applied to obtain the final classification result.
Further, in order to accelerate the convergence rate of the convolutional neural network model and reduce the loss function to the minimum, an Adam optimization algorithm is adopted, the algorithm dynamically adjusts the learning rate of each parameter by utilizing the first moment estimation and the second moment estimation of the gradient, and the calculation process is that the first moment and the second moment of the gradient are firstly calculated through the current gradient of the target function; then correcting the first moment and the second moment; finally, parameters are updated according to the obtained bias correction, expectation and learning rate, and the calculation formula is as follows:
(equation 2):
g t -a gradient of the current parameter;
β 1 /β 2 first order/second order moment attenuation coefficient, i.e. gradient g t /g t 2 (iii) a desire;
m t /v t gradient g t First/second order moments of;
ω t -parameters to be solved (updated);
α -learning rate;
ε=10 -8 ;
t is the number of steps of the update.
Furthermore, the bird repelling module comprises three bird repelling modes of green light bird repelling, sound wave bird repelling and ultrasonic wave bird repelling, the sound wave bird repelling is composed of an MP3 decoding module, a power amplifier and a loudspeaker, the MP3 decoding module is internally provided with an SD card, the input end of the SD card is connected with a microcontroller STM32F, and the output end of the SD card is connected with the power amplifier; the green light bird repelling device comprises green light bird repelling lamps and a light focusing cup, wherein the light focusing cup enhances the irradiation distance of the green light lamps; the ultrasonic bird repelling device comprises an ultrasonic generator and an ultrasonic horn, wherein the ultrasonic generator generates ultrasonic waves with different frequencies.
Furthermore, the detection and bird repelling range expansion module uses a steering engine to work in a matching mode, the detection angle is changed flexibly, and the detection range is expanded.
Furthermore, the illumination detection module is composed of a photosensitive diode DS, a comparator LM393, a capacitor, a resistor and a light emitting diode, and the anode of the photosensitive diode DS is connected with the non-inverting input end of the comparator. Illumination detection detects illumination parameters to determine day or night.
Further, the communication module BC95 module is connected with the IOT Ali cloud server, and the acquired data are uploaded to the cloud server through the BC95 module.
Further, thing networking remote monitoring and module of patrolling and examining include temperature sensor, humidity transducer, electric quantity monitoring, bird invasion information, and the temperature, humidity, electric quantity, the birds invasion number of times of gathering, drive bird information and upload to high in the clouds server, but end user real-time supervision.
Furthermore, the power supply circuit is provided with a solar charging panel and a storage battery, so that the bird repelling system can be normally powered.
Compared with the prior art, the invention has the beneficial effects that: the utility model provides a bird system is driven to thing networking intelligence based on degree of depth study, use STM32F as the core, doppler radar and camera combination, when Doppler radar detects moving object, give STM32F with information transmission, start the camera and shoot the picture, carry out gauss filtering and binarization processing to the picture through OpenCV computer vision storehouse, the rate of accuracy that the target classification detected has been improved, the convolutional neural network model of utilization design, adopt Adam optimization algorithm, carry out degree of depth study, judge whether birds are driven to invading object, if drive the bird mode for birds are opened, adopt the sound wave to drive the bird daytime, the green glow is exploded to dodge and is driven the bird, the ultrasonic wave drives the bird, it is more humanized to drive the bird mode, thing networking remote monitoring, end user is real-time supervision on PC or cell-phone APP. The device that drives the bird when no birds invasion is dormant mode, has reduced the electric quantity loss, and solar panel and battery combine, are showing and have improved duration, have realized that no external power supply independently drives the bird for a long time, and simple and practical drives bird accurately, has reduced fortune dimension degree of difficulty and cost.
Drawings
FIG. 1 is a module architecture diagram of an Internet of things intelligent bird repelling system based on deep learning according to the invention;
FIG. 2 is a target classification schematic diagram of the deep learning-based Internet of things intelligent bird repelling system based on a Doppler detector and deep learning;
FIG. 3 is a target classification neural network model structure of an Internet of things intelligent bird repelling system based on deep learning;
FIG. 4 is a curve diagram of an Acc curve and a Loss curve of the intelligent bird repelling system of the Internet of things based on deep learning;
FIG. 5 is a schematic diagram of image preprocessing of an OpenCV computer vision library of an Internet of things intelligent bird repelling system based on deep learning according to the invention;
FIG. 6 is a network architecture diagram of an Internet of things intelligent bird repelling system based on deep learning according to the invention;
FIG. 7 is an Aliskiren monitoring data interface of an Internet of things intelligent bird repelling system based on deep learning according to the invention;
FIG. 8 is a work flow chart of the Internet of things intelligent bird repelling system based on deep learning.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
The invention provides a technical scheme that: the intelligent bird repelling system based on the Internet of things is composed of an STM32F microcontroller, a bird detection module based on a deep learning algorithm, a multi-mode bird repelling module, a power supply module, an Internet of things remote monitoring and inspection module, a communication module and a detection and bird repelling range expanding module, wherein the detection module is combined with a camera through a Doppler radar, and the system module is structured as shown in figure 1. The working process and principle are as follows:
when the bird detection module detects that the bird moves to the position near the transmission tower, due to the fact that the birds are various in types, different in body types, different in moving speed and different in feather color, whether the birds approach the transmission tower or not is quickly and efficiently detected as a key link of the bird repelling system, and considering that the situation that the birds fly to the position near the transmission tower is a moving process, the bird repelling system selects an HB100 Doppler detector as a pre-detection means. HB100 transmits a fixed frequency (f 0) microwave signal using a doppler radar and receives a reflected wave signal (frequency f 1), the reflected wave frequency being constant when encountering a stationary object; when a moving object is encountered, the difference frequency of the transmitted wave and the reflected wave is mixed to generate a new low-frequency signal, namely a Doppler signal, and the frequency of the Doppler signal is Doppler frequency. The HB100 amplifies the object movement signal and converts the amplified signal into a square wave signal through a zero-crossing comparator, and transmits the square wave signal to the STM32 microprocessor, and the working principle diagram of the HB is shown in fig. 2. The detection range of the detector is a spherical area of 10 meters, the detector can cover a single iron tower, and signals of the approaching and the departing of a moving object can be accurately detected in the detection area. The detector has extremely low power consumption (35 mA), can be in an ultra-long standby state, is not influenced by factors such as temperature, light, humidity and the like, has strong anti-electromagnetic interference capability, adopts non-contact measurement in the whole detection link, cannot damage a detected object, cannot pollute the environment, and accords with the concept of harmonious symbiosis of human and nature. However, doppler is susceptible to the influence of wind and rain in the environment, and the problem of false alarm is easy to occur, so that when a Doppler detector detects a moving object, a signal is sent to an STM microprocessor, a camera is started to shoot a picture, whether the invaded object is a bird is judged by adopting a target classification algorithm based on deep learning, whether the bird repeller is awakened to repel the bird is judged, and the problem of false alarm of the bird repeller system can be effectively reduced.
Target detection algorithms based on deep learning include fast R-CNN, YOLO, SSD and other algorithms, the mainstream algorithms can solve the classification problem of target detection and can detect object position information, but a network model is complex, the requirement on hardware is high, and great computing power is consumed in the training process. In the system involved in the invention, only the bird detection target (an invading object) needs to be distinguished, the species of the bird does not need to be distinguished, and the specific position of the target is not concerned. From the aspects of saving computing power and economy and practicality, the invention designs a simple and practical convolutional neural network model for classification learning so as to identify whether an invading object is a bird or not.
The classification network model structure designed by the invention is shown in fig. 3, and a convolutional neural network with 6 layers is formed by using 2 convolutional layers, 2 pooling layers and 2 full-connection layers. Considering that birds form and postures of the birds stopping and dwelling and flying in the power transmission line are various, the number of the first convolution layer convolution kernels is 8, so that the capability of the network for identifying different intrusion targets is enhanced. The number of output neurons was 2, i.e. the output was both avian and non-avian categories.
Because the image shot by the camera comprises the background such as sky, power transmission lines, insulators, telegraph poles and the like, the bird occupies a small position in the whole image, if the complete image is deeply learned, the whole generalization capability of the model is weak, the accuracy of target classification detection is low, if the generalization capability of the bird needs to be improved, a large number of training samples are needed, and the cost is increased. Therefore, pre-processing is required before the images are imported into the network model training. And performing Gaussian filtering and binarization processing on the background image and the positive sample image through an OpenCV computer vision library, filtering the background, extracting the position and contour characteristics of an invasive object, performing graph segmentation according to the position and contour characteristics, and introducing a network model for training and testing.
The image after the preprocessing has simple and obvious characteristics, so that the image can be converted into a 32 x 32 three-channel pixel format in the image preprocessing stage, the image can be conveniently sent into a network model for training, and the accuracy of recognition is improved. In fig. 3, the input to the first layer of convolutional layers is 32 × 32 pixels in size, where the number of convolution kernels is 8, the convolution kernel size is 5 × 5, and the step size is 1. The dimensions of the output feature map of the first layer convolutional layer are shown in (equation 1).
(equation 1):
o-size of output image;
i-size of input image;
k-the size of the convolution kernel;
p-the number of padding;
s-moving step length.
Since this convolutional neural network does not use all zero padding, P =0. Then, from equation (1), the output of the first layer convolutional layer is 8 × 28 × 28,8 is the number of output feature maps, 8 convolutional kernels output 8 feature maps, and 28 × 28 is the size of the output feature map. The output of the first layer of convolutional layers is taken as the input of the second layer of pooling layers, the filter size of the second layer of pooling layers is 2 × 2, the step size is 2, and the output of the second layer of pooling layers is 8 × 14 × 14. In the third convolutional layer, there are 32 convolutional kernels, the size of the convolutional kernel is 5 × 5, the step size is 1, and the output characteristic map is 32 × 10 × 10. Obtaining 32 multiplied by 5 output through the last layer of pooling layer (the size of the filter is 2 multiplied by 2, and the step length is 2), then connecting with a full connection layer to spread tensor data into vector data with the size of 1 multiplied by 120, and obtaining a final classification result through the full connection layer and applying a softmax activation function.
In order to accelerate the convergence speed of the model and reduce the loss function to the minimum, an Adam optimization algorithm is adopted. The algorithm dynamically adjusts the learning rate of each parameter by utilizing the first moment estimation and the second moment estimation of the gradient, and achieves the purpose of minimizing the loss function. Firstly, calculating a first moment and a second moment of a gradient through the current gradient of an objective function; then, the first order moment and the second order moment are corrected, and the initial value of the parameter is 0, so that the parameter is biased to 0, and the influence of the bias on parameter updating is reduced; finally, parameters are updated according to the obtained bias correction, expectation and learning rate, and the calculation formula is as follows:
(equation 2):
g t -the gradient of the current parameter;
β 1 beta 2-first order/second order moment attenuation coefficient, i.e. gradient g t /g t 2 (iii) a desire;
m t /v t gradient g t First/second order moments of;
ω t -parameters to be solved (updated);
α -learning rate;
ε=10 -8 ;
t is the number of steps of the update.
When the data set is established, about 800 images containing bird targets are collected through various ways aiming at the practical application background of bird target classification detection of the power transmission line, and 200 images without bird targets are added into the data set in order to enhance the robustness of the model. Finally, 1000 images are available for training and testing the model, 80% of which are used as the training set and 20% of which are used as the validation set. The training process is completed on a single personal computer with 2.4GHz and 12GB memory, and a compiler of Python3.7 and a deep learning framework of Tensorflow2.1 are adopted in the training process. The BatchSize was set to 32,epoch =70, and the Acc curve and the Loss curve of the training process are shown in FIG. 4. As can be seen from fig. 4, the accuracy and the loss function of the training set and the verification set have fast rising and falling speeds when the Epoch is 0 to 40, and the curve becomes gentle after Epoch = 60. Finally, the accuracy of the model training set is 98.93%, the accuracy of the verification set is 96.88%, and the requirement of detection precision is met.
After the model is trained on the server, the model can be transferred to an STM32 single chip microcomputer to operate.
In the picture shown in fig. 5 (a), a bird is parked on a power line, and preprocessing is required before the bird is introduced into the network model learning. The background image and the positive sample image are subjected to gaussian filtering and binarization processing through an OpenCV computer vision library as shown in fig. 5 (b), the background is filtered out, and only the positions and contour features of the invasive objects in the images are shown in fig. 5 (c). And (c) projecting the image in fig. 5 (c) along the horizontal direction and the vertical direction respectively to obtain the positions of the vertical and horizontal coordinates of the black pixel point distribution, namely the positions of the intrusion object which can be distributed, as shown in fig. 5 (d) and (e). According to the vertical and horizontal coordinates of the black pixel distribution shown in fig. 5 (d) and (e), the image areas where the intrusion object may be distributed are the areas 1, 2, 3 and 4 marked in fig. 5 (f). The images obtained by clipping the regions 1, 2, 3, and 4 in fig. 5 (f) are respectively introduced into model learning, and it is clear from fig. 5 (f) that only the background is present in the images obtained by clipping the regions 1 and 4, and that only the intruding objects are present in the images obtained by clipping the regions 2 and 3. And (4) outputting a result according to model learning, and acquiring whether birds invade and the invasion times.
The technology based on the narrowband internet of things (NB-IOT) is the most promising technology at present. Bird repellent equipment thing networking is applied generally towards the comprehensive perception to information such as humiture, camera information, electric quantity detection around transmission line and transmission tower, with the information upload to the high in the clouds of gathering, and the end user carries out analysis processes and then realizes functions such as bird pest remote monitoring, patrols and examines and early warning through to high in the clouds information, and thing networking platform framework is as shown in fig. 6. The intelligent bird repeller communication module adopts a remote NB-IOT module BC95 for connecting an NB-IOT base station and supporting various protocols (UDP/TCP/CoAP/LWM 2M/MQTT) to upload data to a cloud platform (China Mobile OneNet, china telecom IoT platform, huashi cloud and Ali cloud). The bird repeller system publishes/subscribes information to the Aliskian cloud based on an MQTT protocol (message queue telemetry transmission), and an Aliskian cloud monitoring data interface is shown in figure 7.
At present, common bird repelling modes include sound wave bird repelling, ultrasonic wave bird repelling, chemical bird repelling and colored light bird repelling. Because the frequency range that birds can hear is narrow, birds can not hear the ultrasonic wave with the frequency higher than 20 khz; and the birds close to the iron tower are driven by chemical drugs, so that the health of the birds is influenced and the environment is possibly polluted. The bird repelling system is used for repelling birds by means of sound waves (voice) and colored light. The eyes of birds are sensitive and can sense colored light with different wavelengths. Of which the most sensitive is to colored light of a wavelength of 500-570 nm. And the red light can influence the migration of the birds and is not in line with the aim of friendly bird repelling, so the bird repelling system adopts green light (520-525 nm) to repel the birds, and the light focusing cup enhances the irradiation distance of green light lamp beads. The bird repeller is a device for repelling birds by voice, and the bird repeller can repel birds by playing different sounds such as the sound of a natural enemy, similar grizzling, gunsound and the like. The sound wave bird repelling module consists of an MP3 decoding module, a power amplifier and a loudspeaker. And the ultrasonic bird repelling module consists of an ultrasonic transmitter and an ultrasonic horn. Drive the bird through changing ultrasonic frequency and then realizing the frequency conversion, can effectively avoid driving the problem of bird effect because of bird's suitable for frequency ultrasonic wave can't realize.
The bird repelling system has a work flow chart as shown in fig. 8, and the work flow is as follows:
(1) an HB100 Doppler detector circularly detects whether a moving object approaches to the power transmission iron tower;
(2) when the existence of a moving object is detected, starting a camera to shoot a picture;
(3) judging whether birds invade or not by using a target classification algorithm based on deep learning (a convolutional neural network is built to train bird pictures to classify invaded targets), and awakening the intelligent bird repeller from a sleep mode;
(4) when the illumination sensor is used for detecting that the current illumination is daytime, the STM32 microcontroller adopts a sound wave and green light flashing mode to drive birds, and pictures are collected; at night, bird repelling is carried out by adopting an ultrasonic wave and green light flashing mode;
(5) the STM32 microcontroller is used for changing green light flashing and ultrasonic frequency to realize high-efficiency bird repelling;
(6) the STM32 timer is used for timing for 30 minutes, the temperature and humidity sensors are started, temperature and humidity data are collected, and the electric quantity of the storage battery and bird invasion frequency data are detected;
(7) detecting whether the BC95 communication module works in a ready state and activates a scene, and registering and logging in the Ali cloud;
(8) the data such as the temperature, the humidity, the storage battery electric quantity, the bird invasion times and the bird repelling information which are collected are uploaded to the cloud server through the BC95 communication module, and a terminal user can monitor and analyze the data in real time.
(9) When the electric quantity is too low or the bird invasion frequency is too high, alarm information is sent to the terminal user through a PC or a mobile phone APP.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (10)
1. The utility model provides a bird system is driven to thing networking intelligence based on deep learning which characterized in that: to transmission line bird pest, the internet of things intelligence that develops drives bird system based on degree of depth study, include STM32F microcontroller, detection module, illumination detection module, drive bird module, power module, internet of things remote monitoring and patrol and examine module, communication module, survey and drive bird scope and enlarge the module, detection module includes Doppler radar, camera, contains following step:
s1: combining a Doppler radar and a high-definition camera to pre-detect and photograph an invading object, carrying out Gaussian filtering and binaryzation processing on a background image and a positive sample image through an OpenCV computer vision library, filtering the background, extracting the position and contour characteristics of the invading object, and carrying out graph segmentation processing;
s2: importing the processed picture into a designed convolutional neural network model for deep learning, and judging whether the invading object is a bird or not;
s3: if the invading object is a bird, the intelligent bird repelling system is awakened from the sleep mode, the bird repelling mode is started, and when the current illumination is detected to be daytime by the illumination detection module, the bird is repelled by sound waves and repelled by green-light flashing; the ultrasonic bird repelling and green light flashing bird repelling are adopted at night, and the green light flashing and ultrasonic frequency is changed by utilizing an STM32F microcontroller to realize bird repelling;
s4: STM32F timer is regularly 30 minutes, starts temperature, humidity transducer, gathers humiture data to detect the electric quantity and the bird invasion frequency data of battery, detect whether the BC95 module is worked and activated the scene, the registration logs in thing networking Ali cloud server, uploads the data of gathering to thing networking Ali cloud server, and end user drives bird information and battery electric quantity through looking over in real time on PC or cell-phone APP.
2. The Internet of things intelligent bird repelling system based on deep learning of claim 1, characterized in that: the camera adopts an OV7725 module, the Doppler radar is combined with the OV7725 camera, when the Doppler radar detects a moving object, detection information is transmitted to the STM32F microcontroller, the OV7725 camera is started to shoot pictures, whether the invading object is a bird or not is judged based on a deep learning classification algorithm, and if the invading object is a bird, a bird repelling mode is started.
3. The Internet of things intelligent bird repelling system based on deep learning of claim 2, characterized in that: the target classification algorithm of deep learning identifies whether an invading object is a bird or not through a convolutional neural network model, the model uses 2 convolutional layers, 2 pooling layers and 2 full-link layers, 6 convolutional neural networks are totally used, the number of output neurons is 2, gaussian filtering and binarization processing are carried out on a background image and a positive sample image through an OpenCV (open computer vision library), the background is filtered, the position and contour characteristics of the invading object are extracted, graph segmentation is carried out, then a network model is led in for training and testing, and the size formula of the output characteristic graph of the convolutional layer of the first layer is as follows:
(equation 1):
o-size of output image;
i-size of input image;
k-the size of the convolution kernel;
p-the number of padding;
s-moving step length.
4. The Internet of things intelligent bird repelling system based on deep learning of claim 3, wherein: the output of the convolution neural network model first layer convolution layer is used as the input of the second layer of pooling layer, the output of the second layer of pooling layer is obtained according to the size and the step length of a filter of the second layer of pooling layer, the output characteristic image size is obtained according to the number of convolution kernels, the size of the convolution kernels and the step length of the third layer of pooling layer, the output obtained through the last layer of pooling layer is connected with a full connection layer to tile and expand tensor data into vector data, and the final classification result is obtained through a full connection layer and by applying a softmax activation function.
5. The Internet of things intelligent bird repelling system based on deep learning of claim 4, wherein: in order to accelerate the convergence rate of the convolutional neural network model and reduce the loss function to the minimum, an Adam optimization algorithm is adopted, the algorithm dynamically adjusts the learning rate of each parameter by utilizing the first moment estimation and the second moment estimation of the gradient, and the calculation process is that the first moment and the second moment of the gradient are firstly calculated through the current gradient of a target function; then correcting the first moment and the second moment; finally, parameters are updated according to the obtained bias correction, expectation and learning rate, and the calculation formula is as follows:
(equation 2):
g t -a gradient of the current parameter;
β 1 /β 2 first/second moment attenuation coefficient, i.e. gradient g t /g t 2 (iii) a desire;
m t /v t gradient g t First/second order moments of;
ω t -parameters to be solved (updated);
α -learning rate;
ε=10 -8 ;
t is the number of steps of the update.
6. The Internet of things intelligent bird repelling system based on deep learning of claim 5, characterized in that: the STM32F microcontroller is a core module of the whole bird repelling system, is externally connected with a detection module, an illumination detection module, a bird repelling module, an Internet of things remote monitoring and inspection module, a power supply module, a communication module and a detection and bird repelling range expansion module, the bird repelling module comprises three bird repelling modes of green light bird repelling, sound wave bird repelling and ultrasonic wave bird repelling, the sound wave bird repelling is composed of an MP3 decoding module, a power amplifier and a loudspeaker, an SD card is arranged in the MP3 decoding module, the input end of the SD card is connected with a microcontroller STM32F, and the output end of the SD card is connected with the power amplifier; the green light bird repelling device comprises green light bird repelling lamps and a light focusing cup, wherein the light focusing cup enhances the irradiation distance of the green light lamps; the ultrasonic bird repelling device comprises an ultrasonic generator and an ultrasonic horn, wherein the ultrasonic generator generates ultrasonic waves with different frequencies.
7. The Internet of things intelligent bird repelling system based on deep learning of claim 6, wherein: the detection and bird repelling range expanding module uses an ostrich machine to work in a matching way, so that the detection angle is flexibly changed, and the detection range is expanded; the illumination detection module consists of a photosensitive diode DS, a comparator LM393, a capacitor, a resistor and a light emitting diode, wherein the anode of the photosensitive diode DS is connected with the in-phase input end of the comparator.
8. The Internet of things intelligent bird repelling system based on deep learning of claim 7, wherein: communication module chooses the BC95 module for use, and the BC95 module is connected with thing networking Ali cloud server, uploads the data of gathering to thing networking Ali cloud server through the BC95 module.
9. The Internet of things intelligent bird repelling system based on deep learning of claim 8, characterized in that: thing networking remote monitoring and patrol and examine the module and include temperature sensor, humidity transducer, electric quantity monitoring, bird invasion information, the temperature, humidity, electric quantity, the birds invasion number of times of gathering, drive bird information and upload to thing networking Ali cloud server, but end user real-time supervision.
10. The Internet of things intelligent bird repelling system based on deep learning of claim 9, wherein: the power supply module is provided with the solar charging panel and the storage battery, so that the endurance time of the bird repelling system is guaranteed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210932069.2A CN115299428A (en) | 2022-08-04 | 2022-08-04 | Intelligent bird system that drives of thing networking based on degree of depth study |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210932069.2A CN115299428A (en) | 2022-08-04 | 2022-08-04 | Intelligent bird system that drives of thing networking based on degree of depth study |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115299428A true CN115299428A (en) | 2022-11-08 |
Family
ID=83858794
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210932069.2A Pending CN115299428A (en) | 2022-08-04 | 2022-08-04 | Intelligent bird system that drives of thing networking based on degree of depth study |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115299428A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117133020A (en) * | 2023-10-26 | 2023-11-28 | 湖北华中电力科技开发有限责任公司 | Power grid facility anomaly detection and processing method and system based on image big data |
CN117238299A (en) * | 2023-11-14 | 2023-12-15 | 国网山东省电力公司电力科学研究院 | Method, system, medium and equipment for optimizing bird voice recognition model of power transmission line |
CN117930381A (en) * | 2024-03-25 | 2024-04-26 | 海南中南标质量科学研究院有限公司 | Port non-radiation perspective wave pass inspection system based on big data of Internet of things |
Citations (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105184824A (en) * | 2015-09-30 | 2015-12-23 | 重庆师范大学 | Intelligent agricultural bird repelling system and method based on image sensing network |
CN107156102A (en) * | 2017-06-01 | 2017-09-15 | 北京中安航信科技有限公司 | Orientation scarer and its bird repellent method based on target following |
CN108174842A (en) * | 2018-01-31 | 2018-06-19 | 佛山市神风航空科技有限公司 | A kind of scarer |
CN208047858U (en) * | 2018-01-30 | 2018-11-06 | 广州环名科技有限公司 | It is a kind of can be to device that power equipment bird pest is expelled in time |
CN110063326A (en) * | 2019-04-30 | 2019-07-30 | 济南浪潮高新科技投资发展有限公司 | Intelligent bird-repeller method based on convolutional neural networks |
CN110167344A (en) * | 2016-11-11 | 2019-08-23 | 生物生态服务提供商 | System and method for detecting flight animal |
CN110753132A (en) * | 2019-11-28 | 2020-02-04 | 佛山光之瞳电子科技有限公司 | Networking monitoring and preventing system and method for bird damage of transformer substation |
CN110827304A (en) * | 2018-08-10 | 2020-02-21 | 清华大学 | Traditional Chinese medicine tongue image positioning method and system based on deep convolutional network and level set method |
JP6704979B1 (en) * | 2018-12-12 | 2020-06-03 | テックファーム株式会社 | Unmanned aerial vehicle, unmanned aerial vehicle system and unmanned aerial vehicle control system |
CN112056302A (en) * | 2020-10-10 | 2020-12-11 | 苏州电力设计研究院有限公司 | Bird repelling device, system and method for main transformer cross arm |
CN112149579A (en) * | 2020-09-25 | 2020-12-29 | 国网江苏省电力有限公司南通供电分公司 | Improved yolo _ v 3-based bird nest hidden danger detection method |
CN112237183A (en) * | 2020-10-14 | 2021-01-19 | 国网山东省电力公司潍坊供电公司 | Monitored ultrasonic closed-loop bird repelling device and bird repelling method |
CN112289326A (en) * | 2020-12-25 | 2021-01-29 | 浙江弄潮儿智慧科技有限公司 | Bird identification comprehensive management system with noise removal function and noise removal method thereof |
CN112364986A (en) * | 2020-11-26 | 2021-02-12 | 南京医科大学 | Convolutional neural network model sensitive to gradient change |
CN212644416U (en) * | 2020-07-30 | 2021-03-02 | 国网福建省电力有限公司南平供电公司 | Lighting of transformer substation and drive integrative intelligent lamp pole device of bird |
CN112614084A (en) * | 2019-09-18 | 2021-04-06 | 中国科学院沈阳计算技术研究所有限公司 | Method for 3D depth convolution neural network aiming at CT image |
CN112806348A (en) * | 2021-01-27 | 2021-05-18 | 郑州师范学院 | Bird repelling system |
CN113180030A (en) * | 2021-07-01 | 2021-07-30 | 广东电网有限责任公司中山供电局 | Embedded bird recognition system |
CN113239877A (en) * | 2021-06-01 | 2021-08-10 | 平安科技(深圳)有限公司 | Farmland monitoring method based on computer vision and related equipment thereof |
CN113287597A (en) * | 2021-05-20 | 2021-08-24 | 河南天通电力有限公司 | Transmission line initiative bird repellent device based on video is studied and judged |
CN214047282U (en) * | 2020-12-08 | 2021-08-27 | 国网北京市电力公司 | Active detection type bird repelling device |
CN113468984A (en) * | 2021-06-16 | 2021-10-01 | 哈尔滨理工大学 | Crop pest and disease leaf identification system, identification method and pest and disease prevention method |
CN114097764A (en) * | 2021-12-22 | 2022-03-01 | 赵世高 | Intelligent bird repelling system |
CN114242080A (en) * | 2021-11-02 | 2022-03-25 | 国网福建省电力有限公司检修分公司 | Distributed bird repelling method and terminal for transformer substation based on bird voiceprint characteristics |
CN114255135A (en) * | 2021-12-28 | 2022-03-29 | 华东交通大学 | Intelligent network connection farmland protection system based on sound and vibration fusion |
CN114661068A (en) * | 2022-05-17 | 2022-06-24 | 天津滨海国际机场有限公司 | Method and device for driving flying object in airport, electronic equipment and medium |
-
2022
- 2022-08-04 CN CN202210932069.2A patent/CN115299428A/en active Pending
Patent Citations (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105184824A (en) * | 2015-09-30 | 2015-12-23 | 重庆师范大学 | Intelligent agricultural bird repelling system and method based on image sensing network |
CN110167344A (en) * | 2016-11-11 | 2019-08-23 | 生物生态服务提供商 | System and method for detecting flight animal |
CN107156102A (en) * | 2017-06-01 | 2017-09-15 | 北京中安航信科技有限公司 | Orientation scarer and its bird repellent method based on target following |
CN208047858U (en) * | 2018-01-30 | 2018-11-06 | 广州环名科技有限公司 | It is a kind of can be to device that power equipment bird pest is expelled in time |
CN108174842A (en) * | 2018-01-31 | 2018-06-19 | 佛山市神风航空科技有限公司 | A kind of scarer |
CN110827304A (en) * | 2018-08-10 | 2020-02-21 | 清华大学 | Traditional Chinese medicine tongue image positioning method and system based on deep convolutional network and level set method |
JP6704979B1 (en) * | 2018-12-12 | 2020-06-03 | テックファーム株式会社 | Unmanned aerial vehicle, unmanned aerial vehicle system and unmanned aerial vehicle control system |
CN110063326A (en) * | 2019-04-30 | 2019-07-30 | 济南浪潮高新科技投资发展有限公司 | Intelligent bird-repeller method based on convolutional neural networks |
CN112614084A (en) * | 2019-09-18 | 2021-04-06 | 中国科学院沈阳计算技术研究所有限公司 | Method for 3D depth convolution neural network aiming at CT image |
CN110753132A (en) * | 2019-11-28 | 2020-02-04 | 佛山光之瞳电子科技有限公司 | Networking monitoring and preventing system and method for bird damage of transformer substation |
CN212644416U (en) * | 2020-07-30 | 2021-03-02 | 国网福建省电力有限公司南平供电公司 | Lighting of transformer substation and drive integrative intelligent lamp pole device of bird |
CN112149579A (en) * | 2020-09-25 | 2020-12-29 | 国网江苏省电力有限公司南通供电分公司 | Improved yolo _ v 3-based bird nest hidden danger detection method |
CN112056302A (en) * | 2020-10-10 | 2020-12-11 | 苏州电力设计研究院有限公司 | Bird repelling device, system and method for main transformer cross arm |
CN112237183A (en) * | 2020-10-14 | 2021-01-19 | 国网山东省电力公司潍坊供电公司 | Monitored ultrasonic closed-loop bird repelling device and bird repelling method |
CN112364986A (en) * | 2020-11-26 | 2021-02-12 | 南京医科大学 | Convolutional neural network model sensitive to gradient change |
CN214047282U (en) * | 2020-12-08 | 2021-08-27 | 国网北京市电力公司 | Active detection type bird repelling device |
CN112289326A (en) * | 2020-12-25 | 2021-01-29 | 浙江弄潮儿智慧科技有限公司 | Bird identification comprehensive management system with noise removal function and noise removal method thereof |
CN112806348A (en) * | 2021-01-27 | 2021-05-18 | 郑州师范学院 | Bird repelling system |
CN113287597A (en) * | 2021-05-20 | 2021-08-24 | 河南天通电力有限公司 | Transmission line initiative bird repellent device based on video is studied and judged |
CN113239877A (en) * | 2021-06-01 | 2021-08-10 | 平安科技(深圳)有限公司 | Farmland monitoring method based on computer vision and related equipment thereof |
CN113468984A (en) * | 2021-06-16 | 2021-10-01 | 哈尔滨理工大学 | Crop pest and disease leaf identification system, identification method and pest and disease prevention method |
CN113180030A (en) * | 2021-07-01 | 2021-07-30 | 广东电网有限责任公司中山供电局 | Embedded bird recognition system |
CN114242080A (en) * | 2021-11-02 | 2022-03-25 | 国网福建省电力有限公司检修分公司 | Distributed bird repelling method and terminal for transformer substation based on bird voiceprint characteristics |
CN114097764A (en) * | 2021-12-22 | 2022-03-01 | 赵世高 | Intelligent bird repelling system |
CN114255135A (en) * | 2021-12-28 | 2022-03-29 | 华东交通大学 | Intelligent network connection farmland protection system based on sound and vibration fusion |
CN114661068A (en) * | 2022-05-17 | 2022-06-24 | 天津滨海国际机场有限公司 | Method and device for driving flying object in airport, electronic equipment and medium |
Non-Patent Citations (1)
Title |
---|
邓庆绪等: ""物联网中间件技术与应用"", vol. 1, 机械工业出版社, pages: 11 - 12 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117133020A (en) * | 2023-10-26 | 2023-11-28 | 湖北华中电力科技开发有限责任公司 | Power grid facility anomaly detection and processing method and system based on image big data |
CN117133020B (en) * | 2023-10-26 | 2024-01-19 | 湖北华中电力科技开发有限责任公司 | Power grid facility anomaly detection and processing method and system based on image big data |
CN117238299A (en) * | 2023-11-14 | 2023-12-15 | 国网山东省电力公司电力科学研究院 | Method, system, medium and equipment for optimizing bird voice recognition model of power transmission line |
CN117238299B (en) * | 2023-11-14 | 2024-01-30 | 国网山东省电力公司电力科学研究院 | Method, system, medium and equipment for optimizing bird voice recognition model of power transmission line |
CN117930381A (en) * | 2024-03-25 | 2024-04-26 | 海南中南标质量科学研究院有限公司 | Port non-radiation perspective wave pass inspection system based on big data of Internet of things |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN115299428A (en) | Intelligent bird system that drives of thing networking based on degree of depth study | |
CN108416963B (en) | Forest fire early warning method and system based on deep learning | |
Liu et al. | High precision detection algorithm based on improved RetinaNet for defect recognition of transmission lines | |
CN106335646A (en) | Interference-type anti-UAV (Unmanned Aerial Vehicle) system | |
CN106971152A (en) | A kind of method of Bird's Nest in detection transmission line of electricity based on Aerial Images | |
CN112806348B (en) | Bird repelling system | |
CN112101088A (en) | Automatic unmanned aerial vehicle power inspection method, device and system | |
CN110850775A (en) | Environmental data wireless monitoring system for intelligent agriculture and working method thereof | |
CN113468984A (en) | Crop pest and disease leaf identification system, identification method and pest and disease prevention method | |
CN112690270A (en) | High-voltage transmission line pole tower bird-detecting and bird-repelling system and method based on millimeter wave radar | |
CN110367239A (en) | A kind of bird trouble on transmission line control method based on Internet of Things | |
CN111160271A (en) | Indoor chicken farm sick chicken detecting system based on deep learning | |
CN114415695A (en) | Tea garden inspection system based on vision technology and inspection robot | |
KR102002632B1 (en) | Fruit monitoring system and method at the same | |
CN114898238A (en) | Wild animal remote sensing identification method and device | |
CN114184232A (en) | Multi-parameter integrated monitoring system for power transmission line | |
Abdulla et al. | Agriculture based on internet of things and deep learning | |
Hu et al. | Agricultural robot for intelligent detection of pyralidae insects | |
CN209330740U (en) | A kind of early warning system based on machine vision Citrus Huanglongbing pathogen | |
CN116616267A (en) | Crop pest control device and method | |
Uchiyama et al. | Sensor network system for condition detection of harmful animals by step-by-step interlocking of various sensors | |
CN114943984A (en) | Intelligent bird detection and bird repelling method and device based on radar point cloud and visual image | |
CN116972906B (en) | Highway state evaluation method, device, storage medium and collection box | |
CN111709385B (en) | Low-altitude and slow-altitude aircraft target detection system and identity recognition method thereof | |
CN109063690B (en) | Analysis method for targets entering transformer substation environment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221108 |