Saliency Detection And Its Application In Wildlife Monitoring Image

Traditional wild animal detection methods can not effectively deal with the problems like high resolution,rich color and high noise in monitoring images.Image saliency detection methods can quickly and automatically extract the salient region representing the main information,remove redundant image information and provide a new way for the rapid detection and recognition of wildlife monitoring images.This paper focuses on the saliency detection and its application in wildlife monitoring image.In addition,reducing the influence of image resolution,rich color and high noise on image saliency detection is also important work target.The main work of this paper is as follows:1.The data set of wild animal monitoring image was established in this paper.We created a data set of 12 species and 1600 wildlife monitoring images based on collecting devices like Infrared Camera and Wireless Multimedia Sensor Network.And the artificial labeled Ground truth data set was built to make the factors affecting the saliency detection of wildlife images clear.It also provided the research material for saliency detection of wildlife monitoring images.2.A Structure and Histogram-based(SHC)contrast method was proposed in this paper.Based on Histogram-based contrast(HC)method,combined with the methods of image structure extraction,edge detection and location saliency map,a structure and Histogram-based contrast method was proposed.The algorithm achieved the optimization of the image saliency detection and improved the quality of detection.3.The saliency detection so ftware for wildlife monitoring images was designed in this paper based on the MATLAB GUI tool.Multiple saliency detection model was collected in this software.And it integrated kinds of functions like image display,parameter setting and debugging and software control function,realizing the integration of different saliency detection model and different image database for evaluation.MSRA 1000 data set and wild animal monitoring images ware tested with SHC algorithm in the experiment.The experimental results show that the proposed algorithm can achieve the wild animal monitoring image saliency detection.And the evaluation index parameters of Precision-Recall curve and the F-measure value were better than the HC algorithm.The work of this part can be used for reference in image saliency detection and wildlife image detection and recognition.The saliency detection software for wild animals can be used to detect wild animals,simplify its operation process,and to facilitate the study of the saliency detection of wild animal monitoring images.