Environment adaptive sensing small sample endangered animal detection method and system

Abstract

The invention belongs to the field of combination of ecological environment and computer vision, and particularly relates to an environment adaptive perception small sample endangered animal detection method and system. A common animal data set and an endangered animal data set are preprocessed, the preprocessed common animal data set is divided into a plurality of subtasks through task decomposition, the subtasks are used for basic stage training, a basic target detection model is obtained, the preprocessed endangered animal data set is sent to an environment sensing module, and a new sample is obtained. A new sample and an original endangered animal data set are used as input to carry out fine tuning training on an existing basic target detection model, and after a trained target model is obtained, an image or a video is input to obtain a related detection result of the endangered animal. A detection model with good generalization ability cannot be trained; and the detection accuracy is low.

Description

technical field [0001] The invention belongs to the combined field of ecological environment and computer vision, and relates to a target detection method in a small-sample scene such as endangered animals, in particular to a small-sample endangered animal detection method and system for environment adaptive perception. Background technique [0002] Endangered animals are an important group in population research and ecological protection, and activity monitoring is one of the important means of related research, but traditional monitoring methods have problems of low efficiency and high labor costs. Therefore, intelligent means based on location tracking and target detection have become an urgent need. Target detection is one of the important research directions in the field of computer vision, and its main task is to identify and locate the target of interest in the image. In recent years, the accuracy and speed of target detection algorithms based on deep models have bee...

AI Technical Summary

Problems solved by technology

[0004] Although the target detection algorithm based on deep learning has made great progress, there are still the following problems in the existing technology: in the small sample scene of endangered animal target detection with extremely scarce image samples, the algorithm cannot be trained to have good generalization ability The target detection model and the adaptability of the scene are insufficient, resulting in low detection accuracy

Images