This paper presents an implementation of a modified parallel-pyramidal algorithm for efficient image processing and identification. The method involves the creation of a system model that supports the integration of spatial, temporal and network data to form a dynamic pyramidal-hierarchical network. The paper details vector sorting techniques, Gtransformations for modifying vector elements, and a shifting procedure that facilitates efficient data transformation. The procedures described are integrated into a general data processing sequence that involves iterative application of these methods until the final result is achieved. How the algorithm works is shown in the example of laser beam projection analysis.
KEYWORDS: Education and training, Machine learning, Evolutionary algorithms, Object detection, Data modeling, Object recognition, Neural networks, Detection and tracking algorithms, Deep learning, Process modeling
This article addresses the development of an intelligent military aircraft identification system using artificial intelligence, machine learning, and deep self-learning technologies to enhance national security and military efficiency. The system aims to automatically and accurately recognize and classify aircraft in images, offering advantages over traditional methods such as higher productivity, speed, accuracy, and the elimination of human error. The importance of deep learning solutions for threat detection and operational efficiency is emphasized. Modern visual data-based object recognition methods and tools are analysed. The methodology includes collecting and preprocessing data, developing a high-precision recognition system based on Yolov8, annotating objects with Roboflow, and creating training, validation, and testing subsets in the yolo format. The paper details the dataset formation process and presents satisfactory results in fast recognition of military aircraft with high classification accuracy. A comparative analysis of Yolov8, R-CNN, and GPT-4 models shows Yolov8's superiority in prediction accuracy and performance. The article describes the model management system for adjusting hyperparameters, selecting object categories, and initiating the training and forecasting process. Testing results demonstrate Yolov8's optimality for military aircraft identification, achieving accurate target identification in complex situations using advanced deep learning algorithms.
In the article the combined reflectance model based on quadratic and cubic polynomials is discussed. The main characteristics of physically accurate Torrance-Sparrow, Löw models and empirical Blinn, Phong, Schlick models are analyzed. The advantages and disadvantages of the cubic and quadratic Blinn-Phong model approximations are explored. The need in the development of new Blinn-Phong model approximation through combining the quadratic and cubic functions is justified. The cubic model is improved in order to improve the accuracy of Blinn-Phong model approximation in the attenuation zone. The formulas of the improved cubic model coefficients are simplified. The precise and approximated formulas for the calculation of connection point between quadratic and cubic functions are obtained. The productivity gain from the replacing the cubic function by the quadratic function in the glare’s epicenter zone is calculated. The absolute and relative errors of Blinn-Phong model approximation by the quadratic, cubic and the proposed model are compared. Through the visualization of the test figures “Teapot” and “Robot” the advantages of the proposed function usage for increasing the realism of glares formation are shown.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.