Coding for the wireless video sensor networks should be a different issue from the coding for video telephony applications. There are two special features observed in the video sensor networks, limited resources, and low fidelity. Due to these features, coding for this network application requires a new perspective. It is a direction for coding to be of simple structure and not for the entertainment purpose but sufficient for the basic sensing needs. Our coding method is accordingly developed as follows: motion compensation followed by a new idea of shape compensation to replace the conventional residue discrete cosine transform coding. More specifically, our binary images are coded by the motion vectors and the class of morphological filters to achieve a bit rate below 10 kp/s. The resources (bandwidth, power, and hardware) are thus altogether significantly saved. This limited resources approach is the first time particularly investigated for the wireless sensor networks.