fpga-based system on chip and its object detection method

Abstract

The invention discloses an object detection method of a system on chip based on FPGA. The method comprises the following steps: an analog receiver acquiring an analog signal and converting the acquired analog signal to an I-path analog signal and a Q-path analog signal, a data acquisition and analog-digital conversion module separately sampling the I-path analog signal and the Q-path analog signal, then conducting analog-digital conversion to obtain a digital complex signal, then transmitting a receiving identification to a state machine; a DDR2-SDRAM conducting cache on the digital complex signal; the state machine, based on the receiving identification, reading the cached digital complex signal from the DDR2-SDRAM, recording the coordinate of each data in a data matrix; a data processing module acquiring a window of the shape of the Chinese character mi, calculating the mean value of reference data of the data matrix so as to determine the object in the data matrix and the position coordinate of the object, at the same time acquiring data calculation completion identification, and transmitting the data calculation completion identification to the state machine, thus completing data processing.

Description

technical field [0001] The invention belongs to the technical field of radar target detection, and in particular relates to an FPGA-based upper system and a target detection method thereof, which are suitable for detection and tracking of targets in noise, clutter or interference environments. Background technique [0002] The echo signal received by the radar contains not only the target, but also various noises, clutter and interference signals. When a fixed threshold is used for detection, if the threshold is set high, missed detection may occur; if the threshold is set low, there will be Lots of false alarms. In order to improve the performance of radar, various clutter suppression measures are usually used in modern radar signal processing, such as moving target display MTI, moving target detection MTD; however, even if the above methods are adopted, clutter remains. When a large number of false targets are recruited, it is easy to overload the radar receiver, so it is...

AI Technical Summary

Problems solved by technology

[0003] Constant false alarm processing generally cannot improve the SNR, but will cause SNR loss. The smaller the number of reference units, the greater the fluctuation of the average estimate, and the noise fluctuation will increase the probability of false alarm. To maintain a constant false alarm Alarm probability, then the SNR of the output must be increased. The SNR loss that needs to be improved is called the false alarm loss.

[0004] The research found that the noise and clutter in the radar background exist simultaneously in the two-dimensional range and Doppler domains, and the local noise can be considered as a smooth Gaussian random process after removing outliers; two-dimensional plane, or combined sliding window processing on the range-Doppler two-dimensional plane, which utilizes information in two dimensions and is more conducive to target detection; and considering the effectiveness of hardware processing, the average constant False alarm (CA-CFAR) detector to achieve, CA-CFAR detector commonly used range-Doppler two-dimensional constant false alarm processing window selection methods include two, one is a rectangular window, the selected reference unit is relatively small The other is the cross window, the selected reference unit is insufficient, and the detection performance is unstable in the uniform clutter background

Images