Pavement deterioration starts from the day of construction and continues to deteriorate along its service life period. in 2018, ministry of road transport and highway, in a report submitted to the supreme court of india informed the apex court about 9300 deaths and nearly 2500 injuries caused due to potholes between 2015 and 2017. selection of the pavement maintenance treatments are based on the type, severity and causes of the pavement distresses. the current practices of functional evaluation of pavement uses manual observations and measurements, which are time consuming, labor intensive and prone to observational errors. a healthy amount of research on development of automated tools for distress detection and quantification using techniques like image processing and remote sensing have been reported in literature. still, the successful in-field use of these methods is very scarce. this research aims at the development of automated, image-processing based tool for evaluation of the deteriorated pavement to support decision making in assigning proper maintenance treatments and rehabilitation techniques to ensure road user safety and increased service life of pavement. a research framework is divided into four stages. first stage includes the identification of the distressed portion of the pavement and segmenting image into damaged and undamaged portions. for this, the images were preprocessed using binarization and morphological transformations. further, the shape of the distress was identified by canny edge detection. the minimum area rectangle algorithms were utilized to distinguish the distressed area from the undamaged pavement. a new methodology for adaptive thresholding named as "normalized intensity thresholding” is developed based on normalized pixel intensity distribution. in stage ii, geometric features of each detected distressed portions were extracted. projective transformation is used to transform pixel dimensions into field dimensions. in stage iii, the spectral and spatial features of the individual distressed segments of the pavement image are analyzed. pixel intensity and image gradient patterns (directional changes in the intensity or color mode) across the distress are plotted and smoothed using basic exponential smoothing. a random forest algorithm was trained for distress classification based on the patterns observed. the results obtained were validated against the field measured values in stage iv. the results show an accuracy of 87% in detection and classification of pavement distress. the accuracy in measurements acquired from the algorithm is 87% for potholes, 95% for patching, and 82% for raveling. the developed algorithm would be very helpful for pavement management organizations and consequently reduction in subjectivity due to the interpretation by the individual. work is ongoing for classification and measurement of cracks.