After all of the field data are collected and organized so that the time and GPS location are linked to the senor data, it is time to classify the data into useful information. Some data like depth and water quality can be easily grouped into appropriate classes, but the video and side-scan data take some additional work.
The HDSS video is classified by applying a standard classification system for each variable under consideration. The individual classes within each category are project dependent, but the general process for each classification is similar. Prior to classification, the observer is trained on a subset of the videos to ensure valid and reliable classification scores. Each video is watched and one category under consideration is scored. The HDSS Video Coder software version 2 (Parham 2014) is used to facilitate the classification process (Figure 1). This software allows the observer to select the appropriate class and have it tied to the second it occurs in the video. Once the classification is completed for the entire group of videos, an overall spreadsheet containing the video file name, the second at which the category occurred, the class name, and the class code is created and stored within a database. This process is repeated for each individual category to be scored.
Numerous classification variables can be developed for the video or side-scan imagery. Habitat types (Figure 2) are classified from the forward facing geo-video while left and right bank condition (Figure 3) are classified from the video of the left and right bank of the river. Substrate type, percent embeddedness, or bottom type are classified using the underwater video or side-scan imagery. In addition to continuous variables such as bank condition that occur at all sites, it is possible to locate and identify discrete objects such as man-made structures, large woody debris, or tributary junctions.
At the heart of the HDSS approach is the ability to gather a wide range of stream corridor information at consistent spatial resolution. This means that your classification system is rapidly and consistently applied to very long stream segments and that you do not end up with highly detailed measurements at one location with no information gathered for miles in either direction. By using HDSS Video Coder software we can create custom classification systems that allow us to differentiate the important features within a particular system.