CN114821662A - Ecological organism identification method based on PFPNet algorithm - Google Patents
Ecological organism identification method based on PFPNet algorithm Download PDFInfo
- Publication number
- CN114821662A CN114821662A CN202210553103.5A CN202210553103A CN114821662A CN 114821662 A CN114821662 A CN 114821662A CN 202210553103 A CN202210553103 A CN 202210553103A CN 114821662 A CN114821662 A CN 114821662A
- Authority
- CN
- China
- Prior art keywords
- ecological
- pfpnet
- algorithm
- image
- biological
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/55—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Databases & Information Systems (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Library & Information Science (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an ecological organism identification method based on a PFPNet algorithm, which comprises the following steps: initiating an ecological organism identification request, and acquiring a biological image in an ecological environment according to the request; collecting ecological biological characteristics, classifying the collected ecological characteristics, and establishing an ecological characteristic database; carrying out target detection on the collected biological image data through a PFPNet algorithm; and comparing and identifying the predicted biological features of the biological image with feature data in an ecological feature database. The invention fuses the multi-scale characteristic graphs output by the SPP by setting the PFPNet algorithm and utilizing the MSCA module, thereby being beneficial to the detection of small targets and being convenient for improving the recognition rate of the targets.
Description
Technical Field
The invention relates to the technical field of biological identification, in particular to an ecological biological identification method based on a PFPNet algorithm.
Background
The aquatic organism community and the water environment have a complex and complicated mutual relationship and play an important role in water quality change. Different types of aquatic organisms have different adaptability to water body pollution, and some types are only suitable for living in clean water and are called as clean water organisms (or oligozoophorous organisms). Some aquatic organisms live in sewage and are called as sewage organisms. The survival marks of the aquatic organisms indicate the water quality change degree, so the organisms become indexes of water pollution, the water pollution condition can be evaluated through the investigation of the aquatic organisms, a plurality of aquatic organisms are sensitive to water poisoning substances, and the water quality pollution degree can also be judged through the toxicity experiment results of the aquatic organisms. The type of the creature is judged by acquiring and identifying the image of the aquatic creature at present, however, the small target is not conveniently and effectively identified in the existing creature identification process, and the biological characteristic identification rate is poor.
Disclosure of Invention
Based on the technical problems in the background art, the invention provides an ecological biological identification method based on a PFPNet algorithm.
The invention provides an ecological organism identification method based on a PFPNet algorithm, which comprises the following steps:
s1, initiating an ecological biological identification request, and acquiring biological images in the ecological environment according to the request;
s2, collecting ecological biological characteristics, classifying the collected ecological characteristics, and establishing an ecological characteristic database;
s3, carrying out target detection on the acquired biological image data through a PFPNet algorithm;
s31, firstly, selecting a biological image, firstly extracting features through a basic network, supposing that the number of extracted feature channels is D, obtaining feature maps with different scales through an SPP network based on the extracted features, expressing the number of scales by N, expressing the number of the obtained feature map channels by CH, wherein CH is equal to D;
s32 then reduces the channels of the feature map obtained in the above step by a channel reduction operation, where the number of reduced channels is represented by CL, and the formula is CL ═ D/(N-1);
s33, obtaining a fused feature map through MSCA operation, wherein the obtained feature channel number is Cp;
s34, finally, forecasting is carried out based on the fused feature maps;
s4 compares the predicted biological features of the biological image with the feature data in the ecological feature database.
Preferably, the biological image selected in step S31 is subjected to CNN feature extraction, and then a feature map of three scales is output through an SPP network.
Preferably, in step S32, the number of channels is reduced through the bottleeck layer, and feature maps of low channel numbers are obtained in three scales respectively.
Preferably, in step S33, the fused feature map is obtained through MSCA operation: MSCA is to perform concat operation on the channel.
Preferably, the number of channels of the fused feature map is: ch + (N-1) Cl ═ D + (N-1) D/(N-1) ═ 2D.
Preferably, in step S1, the biological image is collected as a color image, and the color image of the biological image is converted from an RGB color space to an HIS color space, so as to obtain HIS image information, where the HIS image information includes image saturation, brightness, and position information.
Preferably, the biological image collected in step S1 is subjected to dimension reduction processing, which effectively removes redundant information and extracts useful features.
Preferably, the step S1 processes the acquired biological image data by using an image normalization method and an image enhancement method.
According to the ecological biological identification method based on the PFPNet algorithm, the PFPNet algorithm is set, and the MSCA module is used for fusing the multi-scale characteristic graph output by the SPP, so that the detection of small targets is facilitated, and the identification rate of the targets is improved conveniently.
Drawings
FIG. 1 is a flow chart of the ecological biological identification method based on PFPNet algorithm in the invention;
fig. 2 is a target detection flow chart of the ecological organism identification method based on the PFPNet algorithm provided by the invention.
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.
Referring to fig. 1-2, the ecological biological identification method based on the PFPNet algorithm includes the following steps:
s1, initiating an ecological biological identification request, and acquiring a biological image in the ecological environment according to the request;
s2, collecting ecological biological characteristics, classifying the collected ecological characteristics, and establishing an ecological characteristic database;
s3, carrying out target detection on the acquired biological image data through a PFPNet algorithm;
s31, firstly, selecting a biological image, firstly extracting features through a basic network, supposing that the number of extracted feature channels is D, obtaining feature maps with different scales through an SPP network based on the extracted features, expressing the number of scales by N, expressing the number of the obtained feature map channels by CH, wherein CH is equal to D;
s32 then reduces the channels of the feature map obtained in the above step by a channel reduction operation, where the number of reduced channels is represented by CL, and the formula is CL ═ D/(N-1);
s33, obtaining a fused feature map through MSCA operation, wherein the obtained feature channel number is Cp;
s34, finally, forecasting is carried out based on the fused feature maps;
s4 compares the predicted biological features of the biological image with the feature data in the ecological feature database.
In the invention, the biological image selected in step S31 is subjected to CNN feature extraction, and then a feature map with three scales is output through an SPP network.
In the invention, in step S32, the number of channels is reduced through the bottleeck layer, and feature maps of low channel numbers are obtained in three scales respectively.
In the invention, step S33 obtains a fused feature map through MSCA operation: MSCA is to perform concat operation on the channel.
In the invention, the number of channels of the fused feature map is as follows: ch + (N-1) Cl ═ D + (N-1) D/(N-1) ═ 2D.
In the invention, step S1 acquires a biological image as a color image, converts the color image of the biological image from RGB color space to HIS color space, and obtains HIS image information, which includes image saturation, brightness, and position information.
In the invention, the biological image acquired in the step S1 is subjected to dimension reduction processing, redundant information is effectively removed by the dimension reduction processing, and useful characteristics are extracted.
In the present invention, step S1 processes the acquired biological image data by using an image normalization method and an image enhancement method.
The invention comprises the following steps: initiating an ecological biological identification request, and acquiring a biological image in an ecological environment according to the request; collecting ecological biological characteristics, classifying the collected ecological characteristics, and establishing an ecological characteristic database; carrying out target detection on the collected biological image data through a PFPNet algorithm; firstly, selecting a biological image, extracting features through a basic network, supposing that the number of extracted feature channels is D, obtaining feature maps with different scales through an SPP network based on the extracted features, using N to represent the number of scales, and using CH to represent the number of the obtained feature channels, wherein CH is equal to D; then, reducing the channels of the feature map obtained in the step by a channel reducing operation, wherein the number of the reduced channels is represented by CL, and the formula is that CL is D/(N-1); then obtaining a fused feature map through MSCA operation, wherein the obtained feature channel number is Cp; finally, prediction is carried out based on the plurality of fused feature maps; and comparing and identifying the predicted biological features of the biological image with feature data in an ecological feature database.
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 (8)
1. The ecological biological identification method based on the PFPNet algorithm is characterized by comprising the following steps:
s1, initiating an ecological biological identification request, and acquiring a biological image in the ecological environment according to the request;
s2, collecting ecological biological characteristics, classifying the collected ecological characteristics, and establishing an ecological characteristic database;
s3, carrying out target detection on the acquired biological image data through a PFPNet algorithm;
s31, firstly, selecting a biological image, firstly extracting features through a basic network, supposing that the number of extracted feature channels is D, obtaining feature maps with different scales through an SPP network based on the extracted features, expressing the number of scales by N, expressing the number of the obtained feature map channels by CH, wherein CH is equal to D;
s32 then reduces the channels of the feature map obtained in the above step by a channel reduction operation, where the number of reduced channels is represented by CL, and the formula is CL ═ D/(N-1);
s33, obtaining a fused feature map through MSCA operation, wherein the obtained feature channel number is Cp;
s34, finally, forecasting is carried out based on the fused feature maps;
s4 compares the predicted biological features of the biological image with the feature data in the ecological feature database.
2. The ecological biometric identification method based on the PFPNet algorithm as claimed in claim 1, wherein the biometric image selected in step S31 is subjected to CNN feature extraction, and then is output three-dimensional feature maps through SPP network.
3. The ecological biometric identification method based on the PFPNet algorithm as claimed in claim 1, wherein said step S32 is to reduce the number of channels through bottleeck layer, and three scales respectively obtain feature maps with low channel number.
4. The ecological biometric identification method based on the PFPNet algorithm as claimed in claim 1, wherein said step S33 is implemented by MSCA operation to obtain the fused feature map: MSCA is to perform concat operation on the channel.
5. The ecological biometric identification method based on the PFPNet algorithm, wherein the number of channels of the fused feature map is as follows: ch + (N-1) Cl ═ D + (N-1) D/(N-1) ═ 2D.
6. The PFPNet algorithm-based ecological biometric identification method according to claim 1, wherein the step S1 is to acquire the biometric image as a color image, convert the color image of the biometric image from RGB color space to HIS color space, and obtain HIS image information, which includes image saturation, brightness and position information.
7. The ecological biometric identification method based on the PFPNet algorithm as claimed in claim 1, wherein the biometric image collected in step S1 is subjected to dimension reduction processing, which effectively removes redundant information and extracts useful features.
8. The ecological biometric recognition method based on the PFPNet algorithm of claim 1, wherein the step S1 processes the collected biometric image data by using an image normalization method and an image enhancement method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210553103.5A CN114821662A (en) | 2022-05-20 | 2022-05-20 | Ecological organism identification method based on PFPNet algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210553103.5A CN114821662A (en) | 2022-05-20 | 2022-05-20 | Ecological organism identification method based on PFPNet algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114821662A true CN114821662A (en) | 2022-07-29 |
Family
ID=82516846
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210553103.5A Pending CN114821662A (en) | 2022-05-20 | 2022-05-20 | Ecological organism identification method based on PFPNet algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114821662A (en) |
-
2022
- 2022-05-20 CN CN202210553103.5A patent/CN114821662A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111259786B (en) | Pedestrian re-identification method based on synchronous enhancement of appearance and motion information of video | |
| CN108875608B (en) | Motor vehicle traffic signal identification method based on deep learning | |
| CN109684922B (en) | Multi-model finished dish identification method based on convolutional neural network | |
| CN109410184B (en) | Live broadcast pornographic image detection method based on dense confrontation network semi-supervised learning | |
| CN111582178B (en) | Vehicle weight recognition method and system based on multi-azimuth information and multi-branch neural network | |
| CN109635726B (en) | Landslide identification method based on combination of symmetric deep network and multi-scale pooling | |
| CN112329662B (en) | Multi-view saliency estimation method based on unsupervised learning | |
| CN112464983A (en) | Small sample learning method for apple tree leaf disease image classification | |
| CN114782948B (en) | Global interpretation method and system for cervical fluid-based cytological smear | |
| Khan et al. | Comparitive study of tree counting algorithms in dense and sparse vegetative regions | |
| CN117275048A (en) | Fingerprint identification method based on fusion of global features and local minutiae features | |
| CN116563603A (en) | Small sample open set identification method based on global-local energy collaborative characterization | |
| CN114898464B (en) | Lightweight accurate finger language intelligent algorithm identification method based on machine vision | |
| CN114821662A (en) | Ecological organism identification method based on PFPNet algorithm | |
| CN117523626A (en) | Pseudo RGB-D face recognition method | |
| CN114399793A (en) | Ecological biological identification method based on SPP-Net algorithm | |
| CN114140698A (en) | Water system information extraction algorithm based on FasterR-CNN | |
| CN114565830A (en) | Ecological organism identification method based on CopleNet algorithm | |
| CN114821666A (en) | Ecological organism identification method based on Pelee algorithm | |
| CN115205888A (en) | Ecological organism identification method based on AmoebaNet + NAS-FPN + AA algorithm | |
| CN113159182B (en) | Agricultural tiny pest image detection method based on dense region re-refining technology | |
| CN116863316A (en) | Footprint cross-modal retrieval method based on pressure information | |
| CN114693980A (en) | Light-Head R-CNN algorithm-based ecological biological identification method | |
| CN115147708A (en) | Ecological organism identification method based on MLKP algorithm | |
| CN114842309A (en) | Optimization recognition method for familiar target in familiar scene |
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 |