The Cyborg Astrobiologist: Compressing Images for the Matching of Prior Textures and for the Detection of Novel Textures

摘要

Abstract We describe an image -compression technique of Heidemann and Ritter [4] that is capable of: (i) de -tecting novel textures in a series of images, as well as of: (ii) alerting the user to the similarity of a new image to a previously -observ ed texture. This image -compression technique has been implemented and tested using our Astrobiology phone -cam system, which employs Bluetooth communication to send images to a local netbook server in the field for the image -compression analysis. By providi ng more advanced capabilities for similarity detection and novelty detection, this image -compression technique could be useful in giving more scientific autonomy to robotic planetary rovers, and in assisting human astronauts in their geological exp loration. 1. Introduction In prior work, we have developed computer algorithms for real -time novelty detection and rarity mapping for astrobiological and geological exploration [1 -3,5 -8]. These algorithms were tested at astrobiological field sites using mo bile computing platforms ± originally [5 -7] with a wearable computer connected to a digital video camera, but more recently [1 -3,7 -8] with a phone camera connected wirelessly to a local or remote server computer. The image features used in the novelty dete ction and rarity mapping in prior work were based only upon RGB or HSI color. Nonetheless, even with image features limited to color, the mobile exploration systems worked very UREXVWO\7KHµFRORU -RQO\¶&\ERUJ$VWURELRORJLVWZDVable to identify novel or uncommon areas of image sequences in very different desert environments, ranging from mostly white -colored gypsum to mostly red -FRORUHG µUHGEHG¶ VDQGVWRQHV 7KH V\VWHP ZDVable to identify, for example, lichens of varying colors within the deser t landscapes as being novel features (when first observed) of those landscapes [2-3,7-8] . Herein, we implement and test an image -compression technique of Heidemann and Ritter [4] that is capable of (i) detecting novel (colored) textures in a series of images as well as of (ii) alerting the user to the similarity of a new image to a previously -observed texture. Such a capability could be useful in giving more scientific autonomy to robotic planetary rovers, and perhaps in assisting human astronauts in their ge ological exploration. For example, suppose a semi -autonomous planetary rover equipped with texture -based novelty detection is observing a long series of textures corresponding to hematite con cretions em -bed ded in mineral deposits . Wi th texture -based novelty detection, t his rover would report that a particular previously -unobserved horizontal ly-lay er-ed texture is novel, and hence merits further invest -igation.