Diabetic foot ulcers (DFU) are a serious complication of diabetes mellitus that can lead to lower limb amputations if not detected early. A technique known as foot thermography was employed to locate regions on the feet where ulcers are present by detecting variations in temperature. Studies have revealed that foot thermograms are divided into four categories based on their respective Thermal Change Index (TCI) values, using machine learning algorithms with the k-clustering feature and a convolutional neural network (CNN). In the present work, Linear Discriminant Analysis (LDA) Dimensionality Reduction is used to remove unwanted features and SVM, KNN, Logistic Regression, and Random Forest machine learning models are used for the classification along with Histogram of gradients (HOG) feature extraction. By balancing the dataset through data augmentation, the poor classification was eliminated, and with LDA implementation all the models achieved high accuracy, making it a useful real-time tool.