Uncontrolled garbage disposal related to urban settings can be extremely harmful to the health of individuals as well as the environment. This paper proposes the construction of an intelligent surveillance system based on the YOLOv5 object detection model and an ultra-convolutional net (U-Net), which we will call (YOLOv5-U-Net model), capable of monitoring waste management facilities in real-time through image processing and artificial intelligence. An illustration of an intelligent surveillance system is provided in the following statement. Besides identifying different categories of garbage and possible risks to public health, the system can also identify situations of unsuitable accumulation of waste. For that purpose, to ensure that local authorities will act in due time, the framework integrates several technologies, including object detection algorithms, classification networks as well as real-time warning systems. It is through the amalgamation of these technologies. Testing of the prototype has elicited an outcome; that accuracy related to waste categorization has increased, while reaction times have decreased, all discovered due to prototype examination. Implementation of this strategy does not only increase the linkage between environmental monitoring as well as protection of public health but also gives some help in promoting a conscious urban development, right through ensuring health.