The degree of discernible detail or resolution of an image depends on two digital image parameters. The first parameter is the number of samples in each world dimension, and the second is the number of intensity levels. The larger the number of these parameters the closer the digital image approximates the real world image. Note that the process of developing measures of image fidelity is subjective and application dependent.
The spatial resolution of an image is associated to its number of rows (height) per world dimension and number of columns (width) per world dimension. This also it determines the number of samples required for a "good" approximation.
Image dimensions denotes the number of rows and columns. Most of the time we tend to work with square images where the height and width is a power of two (ex. 128, 256, 512, 1024). Note that these images may or may not have the same resolution.
The number of binary digits, "bits", used in representing the intensity values or grey levels is associated to image quantization. The digitization process of the spatial coordinates is known as image sampling, and the amplitude conversion of an intensity value to a digital representation is know as grey-level quantization.
There is an interesting phenomenon known as the checkerboard effect. This effect is observed by leaving unchanged the number of grey levels and varying the spatial resolution while keeping the display area unchanged. The checkerboard effect is caused by pixel replication, that is, lower resolution images were duplicated in order to fill the display area.
This phenomenon can be visualized in the images shown below
...
------128x128 pixels---------------64x64 pixels
...
-------32x32 pixels---------------16x16 pixels