A Comprehensive Review of Real-Time Image Processing for Industrial Automation Monitoring

Image processing has evolved to be used as a fundamental enabling technology to support industrial automation monitoring, enhance productivity, safety, and decision-making in contemporary manufacturing systems. The rapid adoption of the computer vision and artificial intelligence into the industrial environment has resulted in an accumulating body of research, but the current research is still scattered across the fields of application. The proposed systematic literature review will streamline the recent developments in real-time image processing to monitor industrial automation and also determine the current research trends and performance. To provide rigor and transparency in the review, it is done according to Preferred Reporting Items of Systematic Reviews and Meta-Analyses (PRISMA) protocol. The Web of Science and Scopus databases were searched thoroughly with the use of the advanced search strings, which were developed according to the following keywords: image processing, vision, industrial, automation, and monitoring. Peer-reviewed journal articles with a publication date of 2025 were used in the selection process and 27 primary studies were identified that fit the given inclusion and exclusion criteria. The results were organized and classified into three major themes, namely, (1) vision-based defect detection and quality inspection in industrial production, (2) vision assisted condition, process, and performance monitoring, and (3) vision-based safety, automation, and intelligent industrial systems. Performance indicators in terms of numerical performance reported by reviewed studies portray consistently high detection and monitoring accuracy with many systems having a precision, recall or classification rate of above 95, and real-time processing that is suitable with respect to industrial utilization. In general, the literature review shows that there is a high tendency towards lightweight deep learning models, combination with robotic and cyber-physical systems, and greater focus on real-time safety and autonomous monitoring. The review gives a systematic overview of the existing research findings and notes how image processing methods based on vision have become increasingly mature as important elements of intelligent industrial automation systems.