FUZZY MEASUREMENT OF COORDINATES OF SMALL OBJECTS IN HIGH-RESOLUTION IMAGES

UDC 004.023
DOI: 10.36871/2618-9976.2022.03.004

Authors

Labintsev A.I.
Doctor of Technical Science, Financial University, Moscow, Russia
Dolmatov A.G.
Master of Applied Mathematics, Financial University, Moscow, Russia

Abstract

The article considers the solution of the problem of fuzzy measurement of the coordinates of small objects occupying less than XNUMX% of the area in highresolution images. The EfficientNet and MobileNet classifiers from the Tensorflow library, pre-trained on ImageNet data, are used as the basis of the algorithm for extracting features of small objects. Next, the feature map from the last layers of the neural network is fed to the input of a two-layer perceptron, which implements classification by the feature vector in depth, for each element along the X and Y coordinate axes. The results are interpreted as a measure of the presence of an object in the receptive field in the original image. Thus, the authors have significantly simplified the architecture of the solution, while achieving acceptable indicators of accuracy and precision – XNUMX and XNUMX%, respectively, the inference time on mobile platforms is less than XNUMX second.

Keywords

Object detection, Fuzzy measurements, Image processing, Unmanned aerial vehicles