To reduce the monitoring burden of an observer in an image monitoring device.

An image input part 1 acquires an image signal taken by a monitoring camera or the like, and outputs it to a motion block detection part 2. The detection part 2 divides the image signal inputted from the input part 1 to block units of a fixed size, or divides, for example, a monitoring image of 320240 pixel to blocks of an 8080 pixel unit in this embodiment 1, and determines whether each divided block has motions, based on change of luminance information or the like in the same way as detection of motion vector. A motion block counting part 2 counts the number of motion blocks determined to have motions. An alarm sound control part 4 receives output of the number of motion blocks from the counting part 3, and outputs an alarm sound through a speaker 5 while adjusting the output volume of the alarm sound to be larger as the number of motion blocks is larger.