FIELD: physics.
SUBSTANCE: invention relates to digital image processing. Technical result is achieved by the fact that when obtaining an array of inhomogeneous digital images, priority image processing is carried out in order to reduce the percentage of noise and eliminate signs of distortions. Processing of digital images is carried out sequentially by filtration methods with determination of boundaries of an object on a noisy image, increasing information content. Objects on the image are recognized on the processed digital image by splitting the processed image into streams. In each of the streams, after image convolution, pooling is performed with a filter window of size 3 × 3 pixels with extraction of especially bright features and complete pooling with window size 2 × 2 pixels and reverse process of image restoration to initial input size, after convolution reaches maximum 28 × 28 pixels.
EFFECT: reduction and removal of distortions in an array of non-uniform digital images with increase in information content and high accuracy of detecting target designation objects with further comparison of current images with a reference and distribution by classes.
1 cl, 2 dwg
Title | Year | Author | Number |
---|---|---|---|
METHOD FOR DETECTION AND AUTO-TRACKING OF OBJECTS OF TARGET DESIGNATION BY OPTICAL-ELECTRONIC SYSTEM OF AUNMANNED AERIAL VEHICLE | 2020 |
|
RU2748763C1 |
METHOD AND APPARATUS FOR DETECTING LOCAL FEATURES ON IMAGE | 2013 |
|
RU2535184C2 |
SYSTEM AND METHOD FOR REAL-TIME DATA PROCESSING AND OBJECT RECOGNITION | 2022 |
|
RU2802280C1 |
DEVICE OF SEARCHING IMAGE DUPLICATES | 2013 |
|
RU2538319C1 |
METHOD FOR AUTOMATICALLY RECOGNIZING SCENES AND OBJECTS IN AN IMAGE | 2021 |
|
RU2778906C1 |
METHOD FOR AUTOMATIC CLASSIFICATION OF X-RAY IMAGES USING TRANSPARENCY MASKS | 2019 |
|
RU2716914C1 |
DEVICE TO SEAMLESSLY MERGE THE IMAGES INTO A SINGLE COMPOSITION WITH AUTOMATIC CONTRAST ADJUSTMENT AND GRADIENTS | 2014 |
|
RU2580473C1 |
DEVICE FOR MERGING MEDICAL IMAGES | 2016 |
|
RU2614545C1 |
DATA PROCESSING FOR SUPER-RESOLUTION | 2017 |
|
RU2652722C1 |
METHOD FOR X-RAY TOMOGRAPHY AND APPARATUS FOR REALISING SAID METHOD | 2012 |
|
RU2505800C2 |
Authors
Dates
2022-03-17—Published
2021-03-15—Filed