<![CDATA[Physical violence in the educational environment has a serious impact on mental health, safety, and student achievement, in addition to causing physical injury, violence can cause psychological trauma that interferes with the learning process, due to the limited supervision system, lack of officers, and the absence of automatic detection technology. This research aims to design and develop an automatic detection system of physical violence using digital image processing technology. This study uses the Convolutional Neural Network (CNN) method with the stages of digital image collection and labeling, preprocessing, model training, and evaluation using accuracy, precision, recall, and F1-score metrics. The CNN architecture was chosen because it is efficient and accurate, and it supports data augmentation to improve generalization. The dataset was taken from kaggle and primary data at the al-falah huraba Islamic boarding school which consisted of 2000 images which included: 800 images of violence on CCTV of the dormitory room, 500 images of violence simulation of training videos and 500 non-violent images. The results showed that the developed CNN model was able to detect physical violence with an accuracy of above 88%, making it feasible to apply in surveillance camera-based school surveillance systems (CCTV). The system is able to classify images in real-time into two categories: safe and hard. This research contributes to the use of artificial intelligence to support efficient and affordable technology-based education security.]]>
