Abstract: In the development of computer-vision technology， researchers have primarily focused on processing standard RGB images preprocessed using an image signal processor （ISP）. These images are small as well as convenient for use and network transmission； thus， they are widely applied in many typical application scenarios. However， under low-light conditions or extreme imaging environments， these compressed and processed images typically exhibit irreversible degradations such as blurring and quantization， thus resulting in detail loss and limited performance. Hence， increasing attention has been focused on the direct processing of RAW images output by camera sensors. RAW images， which do not require complex ISP processing， exhibit characteristics such as linear response， large bit depth， and lossless compression， which allow them to retain more original sensory information. These features render RAW images highly flexible and capable in low-light， high dynamic range， and complex visual scenarios. In recent years， RAW-image processing technologies have developed significantly， with their applications extending from high-quality image and video acquisition， denoising， and enhancement to computer-vision tasks such as object recognition and scene understanding. Compared with conventional RGB images， RAW images preserve details more accurately and significantly enhance the accuracy and robustness of visual tasks under specific conditions. Furthermore， the development of deep-learning techniques has enabled end-to-end models based on RAW data， where the original signal information in images are used to enhance visual processing performance. This paper systematically reviews the most recent progress in RAW-image processing technologies and discusses their applications in various computer-vision fields. Additionally， future development trends are forecasted， particularly the potential of using RAW-image data in more complex scenarios， thus providing valuable reference and insights for researchers and practitioners in the field.