Knee osteoarthritis (KOA) is a chronic degenerative joint disease. Early detection is crucial for effective treatment. Recent advances in artificial intelligence (AI) and Machine Learning (ML) have provided powerful tools to automate the diagnosis and classification of KOA. This study mainly focuses on the pre-processing phase of KOA classification, applying advanced image enhancement and noise removal techniques to improve X-radiation (X-ray) image quality for subsequent model training. Approximately 8260 images were taken from a publicly accessible dataset and pre-processed using image enhancement techniques such as Contrast Limited Adaptive Histogram Equalization (CLAHE), Histogram Equalization (HE) and gamma correction and noise removal techniques such as Gaussian filter, Median filter and Bilateral filter. Signal-to-noise ratio (SNR) analysis showed that CLAHE and the Median filter outperformed other techniques in enhancing image and reducing noise, making them the most effective pre-processing methods. These findings are not mentioned in any prior studies, the existing studies use standard pre-processing techniques unlike this study which Statistically Analyses (SA) multiple techniques to find the best combination. Even though the model training phase is yet to be conducted, the study develops a strong pre-processing pipeline that is expected to improve feature extraction and classification accuracy. We also explored recent deep learning (DL) approaches for KOA detection and classification, with a particular emphasis on the potential advantages of ensemble models. Preliminary studies suggest that ensemble model should improve performance by combining multiple predictions, reducing overfitting, and enhancing accuracy, robustness, and generalization to unseen data outperforming the exisiting method. This work also summarizes key developments in AI and ML applications for KOA detection, reviewing existing methodologies while discussing the strengths, limitations, and challenges of current models. Additionally, it highlights promising directions for future research in medical field. The future work will be involving edge detection, feature extraction and model evalution to generate the accuracy, precision, recall and F1-score.