The use of new technologies for monitoring and data collection in earth sciences and river engineering has transformed our understanding of river processes, leading to improved management and preservation of these vital resources. Remote sensing, particularly close-range remote sensing, has emerged as a useful tool for acquiring essential data for river studies. It offers the advantage of large-scale, long-term data collection options, enabling researchers to explore hard to access or hazardous areas and providing a wealth of information to enhance decision-making processes. The importance of remote sensing in earth sciences and river engineering lies in its ability to monitor and collect data on various river hydrodynamics and river processes, such as river ice formation, which significantly influence river characteristics. In cold regions, river ice processes affect hydraulics, sediment transport, water quality, and morphology. The application of close-range remote sensing both using aerial and fixed shore-based imagery in river ice monitoring and data collection has facilitated improved insights into these processes, contributing to better river management and the mitigation of potential hazards. This thesis focuses on the development and application of close-range remote sensing techniques to enhance our understanding of river hydrodynamics and river ice processes. This thesis led to novel applications of close-range remote sensing along two axes: river ice detection and quantification and water survey/discharge measurement. Two algorithms for river ice segmentation and river flow estimation based on artificial intelligent techniques were developed and evaluated in the first axe. The first algorithm is IceMaskNet, a novel river ice detection and segmentation algorithm based on an improved version of the Mask R-CNN. The algorithm has been successfully applied to both aerial and fixed shore-based imagery for river ice detection and classification, achieving average detection and segmentation accuracies of 95% and 91% on aerial imagery. Additionally, the algorithm has been adapted for use on oblique shore-based, low-quality image data, with a detection accuracy of 90% and a segmentation accuracy of 86%. IceMaskNet can be used on aerial imagery to generate quantifiable data and provide insights for an extensive portion of a freezing river. It can also be used on shore-based imagery to gather long-term, near-range observation in comprehending river ice processes. The effectiveness of the developed algorithm was demonstrated in a case study on the Dauphin River where ice categories and ice quantities where extracted over four winters. By employing cost-effective trail cameras along the Dauphin River, a vast collection of oblique, shore-based, and low-quality image data were used to extract quantified river ice data. The comprehensive data and insights derived from this extensive database highlight the potential of close-range monitoring to revolutionize our understanding of river ice processes and their impacts on river systems. IceMaskNet was also adapted, and trained over a set of sea ice imagery to produce an algorithm to identify and segment different sea ice types interacting with bridge piers. The second part of this study was devoted to the development of new tools for water survey and discharge measurement, such as surface velocimetry. In recent years, number of image-based surface velocimetry techniques have emerged, utilizing aerial or shore-based imagery for estimating surface velocity and river discharge. While these methods show great potential in supplementing or even replacing traditional river discharge measurements, they come with high operational costs and require significant user expertise to produce high-quality and satisfactory results. In response to this need, we developed RivQNet, a novel river velocimetry scheme that processes close-range non-contact water surface images using artificial intelligence techniques. The proposed method yields accurate and dense spatial distributions of surface velocities, outperforming conventional optical flow methodologies. Moreover this method requires less amount of user input to estimate surface velocity. RivQNet was further validated with common standard measurement methods and compared with conventional optical flow, Large scale Particle Imagery (LSPIV) and Space Time Image Velocimetry (STIV) methodologies, with a significantly higher estimation accuracy than both LSPIV and STIV, with approximately 25% and 15% higher accuracy respectively for LSPIV and STIV. In conclusion, this thesis demonstrates the value of close-range remote sensing in advancing our understanding of river ice processes and hydrodynamics. The development of novel algorithms, such as IceMaskNet and RivQNet, represents a significant contribution to the field of river engineering and water resources management. The comprehensive data and insights derived from the extensive database of oblique shore-based imagery emphasize the significance of long-term close-range monitoring in gaining a better understanding of river ice processes and hydrodynamics. The developed algorithms can be utilized across a range of applications and settings, benefiting water resources researchers, water survey authorities, and industries engaged in environmental and river engineering projects.