Design Of High Dynamic Image Processing System Based On PYNQ-Z2

In the 1970 s,CCD image sensor became the mainstream image sensor by virtue of its low cost and low noise.With the improvement of semiconductor technology,CMOS image sensor began to gradually replace CCD as the mainstream of the market due to its advantages of high integration,low power consumption and high sensitivity.Dynamic range is one of the most important technical indicators of CMOS image sensor.Due to the fact that the dynamic range of traditional CMOS image sensor is smaller than the actual environment,there are problems of overexposure and underexposure.Therefore,many scholars have been studying CMOS image sensor with high dynamic range.At present,the output noise of high dynamic image sensor is high,and the display equipment cannot satisfy the display of high dynamic image.Based on these problems,there are mainly two kinds of high dynamic image processing at present.One is based on software.Due to the high resolution and high dynamic range of high dynamic image,it needs a lot of resources and the computing speed is slow.One is based on hardware,mainly FPGA.The development cycle of FPGA is short,but the flexibility is poor.It is difficult to store high amount of data and realize complex calculation for high dynamic image.In order to solve the above problems,a high dynamic image processing system is designed based on PYNQ-Z2.PYNQ-Z2 is an SOC development board,which performs image processing on the PL side,and mainly implements IP driving such as VDMA on the PS side,and drives the image output to the monitor for display.High dynamic range image sensor output signal in noise level mainly column fixed pattern noise,gauss noise and salt and pepper noise,this paper put forward the corresponding algorithm and the noise was eliminated,and then the filtered image is automatically focused to obtain a clear image,and then the obtained image is subjected to edge processing to obtain the image frame.When outputting the image,the gamma correction in the global tone mapping is used.High dynamic images are compressed to achieve normal display of images on ordinary monitors,and images are output through the HDMI interface.This paper first verifies the determined algorithm based on the Matlab platform.After verifying the correct results,the algorithm is implemented.That is,the algorithm is written in Verilog code,and the Vivado simulation platform is used to simulate and verify each module.Secondly,the obtained algorithm module is packaged based on the Vivado platform,which is convenient for hardware system construction.Then,integrate according to the hardware design process,assign pins to the interface through the constraint file,and write software code on the SDK platform to implement IP driving such as VDMA.Finally,download the generated file into the PYNQ-Z2 development board for on-board testing.It is verified that the output image can be displayed normally on the ordinary display and the noise is basically eliminated.