Plankton image classification method based on different numbers of samples
A plankton and classification method technology, applied in the field of machine learning and image recognition, can solve problems such as heavy workload and long-term fatigue, and achieve the effect of time-consuming, labor-intensive and error-prone
Pending Publication Date: 2021-12-24
山东易华录信息技术有限公司
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] In order to save labor costs, avoid long-term fatigue of professionals, and solve the problems of plankton classification such as heavy workload, the present invention provides a plankton image classification method based on different numbers of samples
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0021] The specific steps of the inventive method are as attached figure 1 As shown, the main processes include:
[0022] (1) Obtain image data of different types of plankton to form a sample library
[0023] Firstly, different types of plankton images are obtained according to the data collected in the underwater camera equipment, and each type of plankton is classified into one category. The organisms are placed in their respective folders, thus completing the sample library construction. In order to illustrate that in step (2), different feature extraction methods will be adopted depending on the sample size, a random sampling strategy is used to select samples, and a total of 2000 plankton images are selected to construct a small number of plankton sample sets, which are divided into 20 types of plankton. , 100 images for each category; a total of 6,000 images of plankton were selected using a random sampling strategy to construct a large number of plankton sample sets, ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a plankton image classification method based on different numbers of samples. The plankton recognition and classification method based on the random forest and the convolutional neural network comprises the steps that firstly, plankton images of different categories need to be obtained, part of the images are extracted from the planktons of different categories to serve as training sets, and when the number of samples in the training sets is small, the training sets are classified; extracting features (including color features, texture features and morphological features) of each image, and combining the features into feature vectors to train a random forest classifier; when the number of samples in the training set is large, each image is sent into a convolutional neural network to extract abstract features to train a neural network classifier; features of unclassified plankton images are extracted, and the classifier is used for recognition and classification, so that corresponding categories can be recognized. Compared with an artificial plankton classification method, the plankton classification method is high in classification accuracy, low in cost and capable of quickly classifying a large number of planktons.
Description
technical field [0001] The invention relates to the technical field of machine learning and image recognition, in particular to a method for classifying plankton images by extracting plankton image features and using random forests and convolutional neural networks. Background technique [0002] Plankton is a very diverse group of basic microorganisms that exists in almost every water body and plays a vital role in the food chain. Considering the importance of plankton, it is very important to study the automatic classification of plankton. necessary. [0003] The number of plankton taken and extracted by the underwater camera is large, there are many types, and the shapes are different. Manual screening and analysis of these plankton images requires professional technicians in the field of biology to analyze and judge, and the workload is very heavy. The labor intensity is high, but the long-term work fatigue of professionals can also lead to artificial classification erro...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06K9/62G06N3/04
CPCG06N3/045G06F18/214G06F18/24323G06F18/253
Inventor 房开民窦海涛刘文静赵燕
Owner 山东易华录信息技术有限公司