<![CDATA[Disability is a term used for someone who is unable to perform certain activities due to limited physiological, psychological, and anatomical conditions. According to the WHO report, about 15% of the world's population has a disability and in 2020 a Bourne researcher noted that the most people with disabilities are blind or blind people as many as 43.3 million people. The large number of people with disabilities, especially people with disabilities, makes the government and scientists try to find tools that aim to meet the needs of people with disabilities in their activities. Several studies have made tools to detect objects using sensors, but they have weaknesses, namely a narrow range, low accuracy and relatively long computation time. In this study, an architectural disturbance detection system was applied, namely going down stairs and climbing stairs with digital images in bright room conditions (101-1000 lux) using the Gray Level Co-occurrence Matrix feature extraction method and the Artificial Neural Network classification method. Tests were conducted to determine the best distance and theta in feature extraction to produce the highest accuracy, the highest accuracy was obtained with a value of 0.9144 at distance = 3 and theta = 90Ëš. For real time test results, the average detection accuracy is 74.88% and the average computation time is 0.2945528 seconds. The integration of the input and output systems of the test results for the detection of floors, stairs going down and stairs going up results in 100% accuracy.]]>
Copyrights © 2022
