Computer vision is becoming increasingly important. However, the quality and stability of solutions depend not only on the proper post-processing technique (e.g., AI) but also on the quality of the initial images. The better the initial hardware selection, the better the data quality, and consequently, the better—and often more cost-effective—the final result.
Ultimately, a proof-of-concept feasibility study should result in detailed recommendations on your final vision solution's requirements.
Usually, a study starts with a collaborative exploration of specific measurement conditions and associated vision challenges and opportunities. Given these, the study evaluates the optimal combination of:
After intermediate tests on samples in our laboratory, we strongly advise evaluating the final proposed technology on-site, wherever that might be. We have a wide range of cameras, lenses, and tripods to do so. Unique, self-developed hardware enables us to capture data in challenging locations over extended periods to validate the ultimate technology selection.
Camera Technology: The most suitable technological solution depends heavily on the type of problem and the boundary conditions of the challenge. Typically, decisions should made concerning the most interesting frequency range to capture information and the level of detail within this region: RGB, UV, infrared (SWIR, MWIR, LWIR), hyperspectral (Vis-NIR, etc.). Additionally, deciding between a full-field or line-scan camera sensor is crucial.
Thanks to our research activities, we have a unique, extensive portfolio of cameras and hardware that can help you decide on the most suitable technology.
Measurement Strategy: A well-thought-out measurement strategy must be devised alongside the camera technology. For example, in thermography, the type of excitation needs to be specified. In other cases, the spatial positioning of the camera or the consideration between transmission or reflection measurements may be significant.
Lens: Lenses are always chosen based on the type of camera technology and the ultimate necessary amplification or broadening of the camera's field of view (micro vs. macro level).
Illumination: Optimizing ambient lighting conditions (frequency, intensity) can amplify desired effects or suppress background noise. Thus, illumination is a critical quality factor.
Post-Processing Techniques: In an initial study phase, it is often already desirable to consider post-processing, such as classical segmentation methods (thresholding) vs. artificial tools, as these influence the required hardware and, consequently, the cost of the ultimate solution.
Cost: The cost of a final solution is often decisive for implementation. We strive to collaboratively weigh quality, reliability, and cost with the client.