Online detection and identification method for rare animal protection

Abstract

The invention relates to an online detection and identification method for rare animal protection. According to the invention, a traditional image processing method is used to carry out first fuzzy classification on an animal image; the recognition range is narrowed, secondary accurate classification is carried out in a smaller range according to the characteristics of specific animal types, a deep learning method is adopted for aquatic animals and birds, a traditional image processing method is adopted for non-birds, the algorithm is simplified through two-time classification, and the recognition speed is increased. The identification process can be simplified as much as possible; an identification result is obtained through multiple simple classifications; the defect that abstract features cannot be extracted by the conventional method is overcome; the defects that a large number of training samples are needed are overcome, the computing power requirement is lowered as much as possible under the condition that the recognition speed and accuracy requirements are met, hardware configuration is lowered accordingly, the method can be used for achieving image processing on front-end equipment, collected image information does not need to be transmitted to a rear-end computer to be processed, and the implementation mode is simpler and faster.

Description

technical field [0001] The invention belongs to the field of image recognition, and in particular relates to a method for identifying rare animal species for a front-end embedded intelligent device. Background technique [0002] Wild animals are an important part of the natural ecosystem, but due to the over-exploitation and utilization of the natural ecosystem by humans, the area of ​​forests, grasslands, and wetlands has decreased, the living space of wild animals has continued to shrink, and eventually the number of some animals has declined sharply , There are fewer and fewer animal species. under such circumstances, [0003] People have begun to take measures to protect rare animals, among which the establishment of a rare animal field monitoring station is an effective means. The monitoring station generally uses infrared cameras to collect data such as front environmental data and images of rare objects, and then transmits the information to the server. The image da...

AI Technical Summary

Problems solved by technology

[0004] In the identification of rare animal species, more and more image-based identification methods are used to identify animal species. Animal species identification methods generally use traditional image classification techniques to extract features such as shape, color, and texture. Classification and recognition, but this method is difficult to extract more abstract features, and it is difficult to accurately distinguish rare animals with indistinct color distinctions, roughly similar shapes, and variable surface texture topologies

This method requires less computing resources and can run on front-end embedded smart devices, but its accuracy in identifying animal species is low and cannot meet the requirements of real-time identification

In recent years, deep learning methods have also begun to be applied to animal species recognition, which can automatically extract the abstract features of animal images, but this method needs to use a large number of high-quality samples for training, requires more computing resources, and some rare There are relatively few samples of animals, which cannot effectively train the neural network, and for species whose shape and color characteristics are obviously different from other types of animals, it is not necessary to use deep learning

This method can meet the requirements of accurately identifying animal species, but building a training database for all animal species requires huge computing resources, which cannot be realized on front-end embedded smart devices

Images