Melexis' latest time-of-flight 3D camera, the MLX75026, has a fully integrated infrared bandpass (IRBP) filter, eliminating the need to include a separate IR filter in the lens or sensor assembly. Presented as the first available, the solution reduces design complexity and cost, while increasing choice sourcing lenses. The AEC-Q100 qualified QVGA time-of-flight sensor targets automotive use cases like driver and in-cabin monitoring, gesture control, safety monitoring, head-up display driver alignment, and industrial applications like automated guided vehicles, entry management, and robotics.
“An IR filter is mandatory in every ToF application to reduce out-of-band infrared light in the operating environment, which could otherwise lower the dynamic range of the sensor,” explained Gualtiero Bagnuoli, Optical sensors marketing manager at Melexis. “A separate filter adds cost and complexity to a design, but by using the MLX75026 with IRBP filter, engineers can avoid both the cost and complexity, and explore the benefits of time-of-flight sensing more easily”.
A non-trivial effort, integrating the correct IR bandpass filter involves adapting to the illumination and accepting light from a wide range of incidence at minimal spectral shift. Melexis claims to be the only ToF sensor manufacturer currently offering the ability to useToF-lenses with high apertures (low f-number). The MLX75026 also has software compatibility with other Melexis Gen 3 ToF sensors. With a 1/4 inch optical format and supporting 940 nm illumination, device configuration is implemented over an I2C interface, with CSI-2 serial data output.
The EVK75026+ uses a Xilinx Zynq Ultrascale+ MPSoC, and quadruples the data bandwidth and frame rate of the earlier EVK75026 evaluation kit, while taking up approximately half the mechanical size. Offering advanced pixel calibration and filtering functions, the solution also offers on-demand support for features like ambiguous range extension, multi-path resolving, and interference rejection. The programming interface, written in C , includes wrappers that allow the kit to be used with both Matlab and Python.