A novel compression and reconstruction algorithm for aerial image sequence

According to the characteristics of aerial image sequence shot by UAV, a novel compression and reconstruction algorithm based on compressed sensing is proposed. The random measurements of each image and the flight parameters are produced at the encoder, and then transmitted to the decoder. At the decoder, a motion estimation model is built based on analyzing the geometric relationship between camera and object, which can effectively reduce the redundancies between images. The de-correlation image is sparser, which can be reconstructed easily and the recovered images have a higher quality. Experimental results show that the proposed algorithm has improved reconstruction performance and consumes less operating time at the encoder, which is more suitable for hardware implementation on UAV.