Cooperative 3D printing (C3DP) is an emerging field in SWARM Manufacturing (SM) and Additive Manufacturing (AM) that allows for larger prints to be separated and printed more efficiently. Traditional 3D printing is constrained by the dimensions of the print bed and 3D printing arm reach. However, C3DP allows parts to be printed using multiple arms, providing a faster, more efficient way to print wider and larger parts. Our previous work in C3DP includes geometric partitioning, scheduling, placement, path planning, and primitive in-situ monitoring. This paper aims to further the in-situ monitoring technology by applying Canny edge detection for better image matching accuracy and retraining the computer vision machine learning model with over 7000 images rather than the original 276 images used in the first iteration of this research. By refining these techniques, the system gains robustness across diverse camera perspectives and challenging error conditions, enhancing its adaptability and reliability in real-world applications. In other words, the focus of this paper is to enhance the error detection model by harnessing Computer Vision-based software image processing techniques such as edge detection and image augmentation. Furthermore, a proper closed-loop system where, as soon as a certain level of stringing or incorrect matching is present, the printer is informed and stops the print, recalibrates, and restarts the printing sequence from that point. Our previous research with in-situ monitoring of print jobs worked with 2D side views and nozzle tracking using ArUco markers from a top-down view. The biggest gap seen in this first iteration is the lack of accurate annotations and image matching scores for the side view error detection causing an ineffective closed-loop system. Furthermore, it is important to note that although previous studies have investigated machine learning models to detect errors in 3D printing and utilizing edge detection for print similarity, this study focuses on the following novelty: in-situ monitoring in a closed-loop feedback system that can confidently stop and recalibrate given the information from an Intel RealSense depth camera and an intricate software processing workflow. By utilizing traditional edge detection techniques to improve the image matching score and a largely trained model, including image augmentation, the goal is to improve the accuracy of detection regardless of various camera perspectives and error conditions. This research will show the improvement from the original model with a side-by-side in-situ analysis, demonstrating the increased effectiveness of real-time detection and adjustment in cooperative 3D printing. The results will highlight the potential of this enhanced system to be adapted across various applications in the additive manufacturing industry, ultimately moving toward autonomous, high-precision 3D printing systems in complex manufacturing environments.