Patent Application
20200066785
The invention relates to a TDI line detector.
Particularly in the field of earth observation, various methods and sensors are known that scan the earth from orbit with the aid of various sensors (lines or matrixes). Using various technologies, the aim is to maximize the resolution of the instrument both spectrally and geometrically. Examples of such technologies are sensor architectures with TDI line detectors or movement-compensated matrix detectors and/or line detectors.
DE 10 2008 024 417 B3 has disclosed a TDI line detector comprising n TDI lines, each TDI line having m pixels. In this case, at least one outer line is associated with a readout register unit, each pixel of the first and/or nth line being associated with a separate register. In it, a masking is provided between two pixels of each line, the masking having a width of one pixel or an integral multiple of one pixel and the registers of the pixels being widened in accordance with the width of the maskings. In addition, a detector device is disclosed, which has at least two TDI line detectors, the respective TDI line detectors being positioned offset from each other by one pixel width in the transverse direction and being positioned offset from each other by an integral multiple of one pixel length in the longitudinal direction. The TDI line detectors in this case are preferably synchronized with each other and the detector device can also be of a monolithic design.
One problem in TDI line detectors with multiple TDI stages is the stability in order to limit the jitter between the pixels to less than ⅓ of a pixel during recording. For example if 1,000 lines are clocked with a clock frequency of 10 kHz, then the line detector must be kept stable for 100 ms in order to comply with the jitter requirements. This particularly limits the desire for an increase in the number of lines. But even with the conventional number of lines, this requires an increased complexity, for example because of so-called microvibrations in a satellite.
The TDI line detector comprises n TDI lines, each TDI line having m pixels. In this case, the TDI line detector is divided into x submodules, the number of lines of a submodule being n/x. In addition, each last line of each respective submodule is associated with a separate set of readout electronics, the length of the readout electronics corresponding to an integral multiple of the length of a pixel, where x≥2. In this case, the associated pixels of different submodules are positioned relative to one another with pixel-level accuracy. This achieves the fact that the time during which the TDI line detector has to be kept stable is reduced by the factor x with the same number of lines. In this connection, it should be noted that the n TDI lines are activated synchronously and the TDI line detector is preferably embodied as a monolithic chip, more preferably using CMOS technology.
The data that are then read out are transmitted via a bus system to an evaluation unit where they are correlated with one another by means of known methods. One possible method, for example, is the center-of-gravity method.
Alternatively, every second submodule or alternating groups of submodules is/are positioned laterally offset by a half-pixel, which makes it possible to increase the geometrical resolution. In this embodiment as well, the stability problem is reduced by the factor x.
In another embodiment, 10≤x≤100. This strikes a very good compromise between mitigation of the stability problem and the additional expense for the readout electronics and postprocessing.
In another embodiment, at least some of the submodules have filters for various spectral ranges so that a multi-spectral sensor can be achieved. This can be combined with panchromatic submodules. In such an embodiment, it is for example provided that if three different spectral ranges are being mapped, then three submodules form a group, i.e. the first three submodules are positioned in a line and the next three submodules are positioned offset laterally by a half-pixel. With the additional integration of a panchromatic submodule, the group would then consist of four submodules.
In another embodiment, the pixel sizes are the same for all lines. Particularly in mixed embodiments in which panchromatic and spectral submodules are mixed, this avoids aliasing effects. The spectral pixels can then be combined in the evaluation unit for example through the formation of macro-pixels in order to achieve an improved signal-to-noise ratio.
In another embodiment, the pixels are rectangular, with the pixels being twice as wide as they are long. Particularly in the embodiment in which individual submodules are positioned laterally offset by a half-pixel in order to increase the geometrical resolution, this therefore yields square pixels again in the postprocessing.
A preferred application field of the TDI sensor is the use in a satellite.
The invention will be explained in greater detail below based on preferred exemplary embodiments. In the figures:
The length L1 of the readout electronics 11-14 in this case is an integral multiple of the length L2 of the pixels P. The associated pixels P are positioned relative to one another with pixel-level accuracy. The lines Z1-Zn and the readout electronics 11-14 are then synchronously clocked. This division into submodules S1-S4 reduces by a factor of 4 the time during which the position of the TDI line detector 1 has to be stable (e.g. deviation in the spatial position of less than ⅓ of a pixel). This can then also be used, for example, to increase the number n of lines since the stability requirements can be easily met by means of a correspondingly high number of submodules.
This is of interest particularly for multispectral recordings since they usually use larger pixel areas.
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2017 208 041.8 | May 2017 | DE | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2018/062095 | 5/9/2018 | WO | 00 |
