Shape adaptive DCT compression for high quality surveillance using wireless sensor networks

Wireless surveillance networks consists of numerous camera and sensor node to transmit the surveillance details from a remote location to the user nodes. Large amount of information transmitted via the sensor nodes are non-priority information like the background which never changes throughout the surveillance time. This non priority information requires more space and is of no use in transmitting so this information can be compressed to the maximum level without affecting the quality of the image transmitted. For efficient transmission of the input data it is important that only Region of Interest (ROI) is transmitted with lower compression ratio and the non ROI regions to be compressed as much as possible. In this proposed work a shape adaptive DCT compression and Decompression scheme is proposed for efficient image data transmission over the wireless sensor networks. The frames from the surveillance sensor networks are acquired and the ROI is calculated using dynamic saliency maps. The image is then divided into two parts, the transmitting node performs the shape adaptive DCT on the image and transmits the image to the user node where the decompression is done using the inverse shape adaptive DCT. The performance of the proposed algorithm is tested on the set of video images and performance is tabulated for the quality of image and current consumption when the compressed image is transmitted by sender ENTDEV019 ESP 8266 WiFi MCU node and received by receiver ENTDEV019 ESP 8266 WiFi node.