Vehicle Detection in Remote Sensing Image Based on Machine Vision.

Liming Zhou,Bing Zhou,Haoxin Yan,Baojun Qiao,Yang Liu,Chang Zheng,Xianyu Zuo,Minghu Fan,Anastasios D Doulamis

doi:10.1155/2021/8683226

Abstract

Target detection in remote sensing images is very challenging research. Followed by the recent development of deep learning, the target detection algorithm has obtained large and fast growth. However, in the application of remote sensing images, due to the small target, wide range, small texture, and complex background, the existing target detection methods cannot achieve people's hope. In this paper, a target detection algorithm named IR-PANet for remote sensing images of an automobile is proposed. In the backbone network CSPDarknet53, SPP is used to strengthen the learning content. Then, IR-PANet is used as the neck network. After the upper sampling, depthwise separable convolution is used to greatly avoid the lack of small target feature information in the convolution of the shallow network and increase the semantic information in the high-level network. Finally, Gamma correction is used to preprocess the image before image training, which effectively reduces the interference of shadow and other factors on training. The experiment proves that the method has a better effect on small targets obscured by shadows and under the color similar to the background of the picture, and the accuracy is significantly improved based on the original algorithm.

Highlights

Remote sensing target detection is to mark the object of interest in remote sensing images and forecast the type and location of this targets
In the traditional detection dataset, the target is concentrated, while the aviation dataset is not, and the object strength in the aviation image usually appears in arbitrary orientation, which depends on the perspective of the Earth vision platform [1]
We propose the IR-PANet algorithm for targets in shadow occlusion and targets with a similar color to their surroundings in remote sensing images. e detector replaces convolution with inverted residual [27] based on PANet, which can increase the depth of the model, enrich the semantic information of the algorithm and increase its detection accuracy

Summary

Introduction

Remote sensing target detection is to mark the object of interest in remote sensing images and forecast the type and location of this targets. The class of vehicle in remotely sensed images is often subject to weather and environmental images such as atmospheric occlusion, shadow occlusion, and building occlusion and other factors, for example, different overhead views, different sizes of vehicle targets in the same image, similar colors between vehicles and their surroundings, and so on. It resulting in poor detection accuracy of car targets

Methods

Results

Conclusion