Search

CN-121996608-A - Wide spectrum image acquisition and recognition system based on FPGA

CN121996608ACN 121996608 ACN121996608 ACN 121996608ACN-121996608-A

Abstract

A set of wide spectrum image acquisition and recognition system based on FPGA belongs to the technical field of photoelectric sensing and edge intelligent computing intersection. The system utilizes the controllable photoelectric characteristic of the quantum dot material to construct a device array, realizes wide spectrum sensing of visible light to near infrared band light signals, and directly introduces multichannel photoelectric signals generated by the array into an FPGA platform near a sensing end for processing. The functional modules such as analog signal acquisition, analog-to-digital conversion, data caching and neural network reasoning are integrated in the FPGA, so that the cooperative operation and integrated processing of optical signal acquisition, data conversion and intelligent identification are realized. Compared with the recognition scheme based on the general processor, the invention fully plays the advantages of parallel computation and reconfiguration of the FPGA, reduces the data transmission cost, improves the speed and reduces the overall power consumption. The system has compact structure and high integration level, and can complete real-time sensing and digital identification of the wide-spectrum optical signals at the edge side.

Inventors

  • WANG XIAOTING
  • SHI HAOLIN
  • CHENG CHUANTONG
  • ZHANG YONGZHE

Assignees

  • 北京工业大学

Dates

Publication Date
20260508
Application Date
20260126

Claims (9)

  1. 1. A set of FPGA image acquisition and recognition system based on a wide spectrum quantum dot detector array is characterized in that an all-inorganic metal halide perovskite (CsPbBr 3 ) and narrow-band-gap lead sulfide (PbS) quantum dot are coupled to construct a 4x 4-scale wide spectrum response detector array, each detection unit corresponds to an independent pixel point of a digital image, an FPGA is used as a main control chip, each column of signals of the array are gated by an analog switch in a time-sharing manner, the gated current signals are converted into voltages through an operational amplifier and amplified, the voltages are digitized by an analog-to-digital converter AD and then stored in the FPGA, the digital signals are recognized in real time based on A Neural Network (ANN) model pre-deployed on the FPGA, finally, recognition results are displayed through a nixie tube, and meanwhile, current data of each unit of the array are uploaded to an upper computer for analysis through a serial port.
  2. 2. The FPGA image acquisition and recognition system based on the wide spectrum quantum dot detector array is characterized by comprising 2 PCBs, wherein one PCB is a main board comprising an FPGA chip, an AD conversion module, a nixie tube display module and a serial port module, and the other PCB is an auxiliary board comprising an analog switch module, an operational amplifier module, a power supply module and a 4x4 quantum dot device array which is pressed and welded on the auxiliary board through a pressing and welding process.
  3. 3. The FPGA image acquisition and recognition system based on the wide spectrum quantum dot detector array is characterized in that an analog switch can control 4 paths of signals and respectively control the reading of electric signals of different devices on a chip, an operational amplifier can realize the operational amplification of the 4 paths of electric signals, an AD conversion chip can realize the analog-to-digital conversion of 8 channels at most, a serial port communication adopts an RS232 interface, and a special power chip is used for supplying power to provide +/-5V.
  4. 4. The FPGA image acquisition and recognition system based on the broad spectrum quantum dot detector array, which is disclosed in claim 1, is characterized in that the structure of the quantum dot device is that a source electrode and a drain electrode are arranged on a silicon oxide wafer, a single-layer graphene film is sequentially arranged on the silicon oxide wafer and between the source electrode and the drain electrode from bottom to top, the single-layer graphene film is used as an electric conduction layer for providing a carrier transmission channel, a PbS-TBAI quantum dot film, wherein TBAI is tetrabutylammonium iodide and is used for carrying out surface ligand exchange on PbS quantum dots to improve electrical transmission performance, a CsPbBr 3 quantum dot film is used for light absorption and generation of photo-generated carriers, and a Parylene film is used as a packaging protection layer.
  5. 5. The FPGA image acquisition and recognition system based on the broad spectrum quantum dot detector array, which is disclosed in claim 4, is characterized in that the source electrode and the drain electrode in the quantum dot device are both Ti/Au, the graphene film is a single-layer graphene film, the single-layer graphene film is transferred onto a silicon oxide wafer in deionized water, the PbS-TBAI quantum dot film and the CsPbBr 3 quantum dot film are prepared by adopting a spin coating method, wherein the number of layers of the PbS-TBAI quantum dot film is 3, the number of layers of the CsPbBr 3 quantum dot film is 1, and the number of layers of the Parylene film is 1.
  6. 6. The FPGA image acquisition and recognition system based on the wide spectrum quantum dot detector array according to claim 1, wherein the neural network model is an ANN neural network, the neural network is composed of 16 neurons in an input layer, 1 layer in a hidden layer and 12 hidden neurons in an output layer and 4 neurons in an output layer, training of the neural network is completed through Matlab software, after training of the neural network is completed and the recognition accuracy reaches more than 95%, weight parameters (weights) and bias parameters (bias) of the neural network are extracted and stored, and then the neural network model containing the weight parameters and the bias parameters is deployed on the FPGA.
  7. 7. The system for acquiring and identifying an FPGA image based on a wide spectrum quantum dot detector array according to claim 1, wherein 16 neurons of the input layer of the neural network represent 16 devices respectively, and 4 neurons of the output layer represent 4 identification results (0,1,4,7) respectively.
  8. 8. The system for acquiring and identifying an FPGA image based on a wide spectrum quantum dot detector array according to claim 1, wherein the digital pixels of the digital identification are 4x4 in size by projecting a 4x4 digital image onto the array of 4x4 devices using a mask such that each device represents a pixel.
  9. 9. The system for acquiring and identifying the FPGA image based on the wide spectrum quantum dot detector array according to claim 1, wherein the FPGA development adopts Vivado, the hardware description language adopts Verilog language, and the PCB drawing adopts Altium Designer.

Description

Wide spectrum image acquisition and recognition system based on FPGA Technical Field The invention relates to the field of FPGA, quantum dot detector arrays, artificial neural networks and edge image acquisition and identification, in particular to an FPGA image acquisition and identification system based on a wide spectrum response quantum dot device array. Background In the sensing fields of automatic driving complex environments and the like, short wave infrared (SWIR, 0.9-1.7 mu m) imaging is carried out by analyzing the reflection/absorption characteristics of a target on photons with specific wavelengths, such as haze, weak light and the like. Exhibiting unique advantages in the environment. However, silicon-based detectors are limited by a 1.12 eV bandgap (cutoff wavelength λ c ≡1100 nm), and the practical upper effective response limit is typically 1000 nm, so conventional silicon technology cannot meet SWIR imaging requirements. And the Quantum Dot (QD) device can regulate and control the response wave band to the SWIR interval through the size-dependent quantum confinement effect. In addition, the traditional SWIR image intelligent processing relies on general processors such as a CPU (Central processing Unit), a GPU (graphics processing Unit) and the like, is limited by separation of Von Neumann architecture calculation and storage, has the problems of higher power consumption, high data transmission cost, insufficient instantaneity and the like when intelligent algorithms such as a neural network and the like are operated, and is difficult to meet the requirements of an edge scene on low power consumption and quick response. In contrast, the FPGA has a highly parallel hardware architecture and good reconfigurable characteristics, can be deployed at the near end of the quantum dot imaging array, and can finish high-speed acquisition and intelligent calculation acceleration of output signals and match low-delay requirements of edge perception. Disclosure of Invention Therefore, the invention provides a set of FPGA image acquisition and recognition system based on a wide spectrum quantum dot detector array, which is used for carrying out high-speed acquisition and intelligent calculation and recognition on the output signals of the detector array. The system comprises a 4x 4-scale wide spectrum response detector array, wherein each detection unit corresponds to an independent pixel point of a digital image, an FPGA is used as a main control chip, each column of signals of the array are gated in a time-sharing mode through an analog switch, the gated current signals are converted into voltages through an operational amplifier and amplified, the voltages are digitized by an analog-to-digital converter AD and then stored in the FPGA, the digital signals are recognized in real time based on A Neural Network (ANN) model pre-deployed on the FPGA, a recognition result is finally displayed through a nixie tube, and meanwhile, current data of each unit of the array are uploaded to an upper computer for analysis through a serial port. The system can realize acquisition, conversion and identification of digital image signals with the size of 4x4 pixels. The invention aims at realizing the system by adopting the following technical scheme, and the FPGA image acquisition and recognition system based on the wide spectrum response quantum dot device array mainly comprises 2 PCBs, wherein one PCB is a main board and comprises an FPGA chip, an AD conversion module, a nixie tube display module and a serial port module, the other PCB is a secondary board and comprises an analog switch module, an operational amplifier module, a power supply module and a 4x4 quantum dot device array which is welded on the secondary board through a pressure welding process. Further, the main control FPGA chip is produced by Xilinx company, the specific model of the chip is XC7A35TFGG484-2, an Analog switch can control 4 paths of signals to respectively control the reading of electric signals of different devices on the chip, the specific model is MAX4634, an operational amplifier is produced by Texas instrument (Texas Instruments), the specific model is OPA4192, the AD conversion chip is produced by Adenox (Analog DEVICES INC), the maximum of 8 channels of Analog-to-digital conversion can be realized, the specific model is AD7606, serial port communication adopts an RS232 interface, a special power chip is used for supplying power for +/-5V, the chip for supplying +5V power supply voltage is produced by LINEAR TECHNOLOGY company, the model is LTC1522, and the chip for supplying-5V power supply voltage is produced by LINEAR TECHNOLOGY company, and the model is LTC1983-5. The quantum dot device is characterized in that a source electrode and a drain electrode are arranged on a silicon oxide wafer, a single-layer graphene film is arranged on the silicon oxide wafer and between the source electrode and the drain electrode in sequence from