Research Repository

Probabilistic Faster R-CNN with Stochastic Region Proposing: Towards Object Detection and Recognition in Remote Sensing Imagery

Yi, Dewei and Su, Jinya and Chen, Wen-Hua (2021) 'Probabilistic Faster R-CNN with Stochastic Region Proposing: Towards Object Detection and Recognition in Remote Sensing Imagery.' Neurocomputing. ISSN 0925-2312

[img] Text
NEUCOM-D-21-01003_R1_accepted.pdf - Accepted Version
Restricted to Repository staff only until 30 June 2022.
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (4MB) | Request a copy

Abstract

Object detection is one of the most important tasks involved in intelligent agriculture systems,especially in pest detection. This paper focuses on a most devastated agricultural disaster: grasshopperplagues. Grasshopper detection and monitoring is of paramount importance in preventing grasshopperplagues. This paper proposes a probabilistic faster R-CNN algorithm with stochastic region proposing,where a probabilistic region proposal network, an image classification network, and an object detectionnetwork are integrated to detect and locate grasshoppers. More specifically, in the proposed framework,the probabilistic region proposal network considers attributes (e.g. size, shape) of region proposals andthe image classification network identifies the existence of grasshoppers while the object detectionnetwork scores recognition confidence for a region proposal. By integrating these three networks, theuncertainty can be passed from end to end, and the final confidence is obtained for each region proposalcan be explicitly quantified. To enhance algorithm robustness, a stochastic region proposing algorithmis developed to screen region proposals rather than using a predetermined threshold. The proposedalgorithm is validated by recently collected grasshopper datasets. The experimental results demonstratethat the proposed algorithm not only outperforms competing algorithms in terms of average precision(0.91), average missed rate (0.36), and maximumF₁-score (0.9263), but also reduces the false positiverate of recognising the existence of grasshoppers in an open field.

Item Type: Article
Uncontrolled Keywords: Object detection, Image recognition, Gaussian mixture models, Region proposal network
Divisions: Faculty of Science and Health > Computer Science and Electronic Engineering, School of
Depositing User: Elements
Date Deposited: 28 Jun 2021 13:56
Last Modified: 02 Jul 2021 20:15
URI: http://repository.essex.ac.uk/id/eprint/30666

Actions (login required)

View Item View Item