閃爍定義

光學(xué)閃爍被定義為人造光源的脈動(dòng)或波動(dòng)的光現(xiàn)象。
在低頻閃爍中，光是可見的（人眼能夠察覺光的閃爍）。超過100 Hz的光學(xué)閃爍對(duì)于人眼來(lái)說不再可見，但仍然存在，可能對(duì)人體產(chǎn)生影響。
大多數(shù)類型的人造光源在連接到電力主網(wǎng)（家庭或商業(yè)辦公室）時(shí)會(huì)發(fā)出閃爍，這主要取決于國(guó)家的電力頻率，通常是50 Hz或60 Hz。
由于電流在光源中的交替流動(dòng)，所有人造光源都會(huì)分別在50 Hz和60 Hz電力主網(wǎng)下發(fā)出100 Hz或120 Hz的閃爍頻率。
為了消除可見的閃爍并減少對(duì)人體的影響，大多數(shù)發(fā)光二極管（LED）使用脈沖寬度調(diào)制（PWM）調(diào)光方法，以實(shí)現(xiàn)更高的閃爍頻率。
VD6283傳感器可以檢測(cè)光的閃爍頻率，最高可達(dá)2 kHz。
最近在弄ST的課程，需要樣片的可以加群申請(qǐng)：615061293 。

視頻教學(xué)

https://www.bilibili.com/video/BV1Fh4y1r7AS/

樣品申請(qǐng)

https://www.wjx.top/vm/OhcKxJk.aspx#px#)

源碼下載

https://download.csdn.net/download/qq_24312945/88394817

開發(fā)板設(shè)置

在手冊(cè)種給出了，閃爍手冊(cè)可以查看AN5639，資料鏈接如下。https://www.st.com/content/ccc/resource/technical/document/application_note/group1/9f/7e/8c/ce/36/85/4c/08/DM00776948/files/DM00776948.pdf/jcr:content/translations/en.DM00776948.pdf

在AN5639手冊(cè)中，需要對(duì)SB3進(jìn)行連接。

同時(shí)GPIO2需要接到MCU的ADC通道中。

查看X-NUCLEO-6283A1手冊(cè)，可以看到VD6283TX的GPIO2連接到MCU的ADC端口0-2。

在VD6283TX-SATEL中，可以看到VD6283TX通過SB3連接到了AFLR_1V8。
需要將AFLR_1V8接到開發(fā)板的A0端口中。

開發(fā)板選擇

這里使用NUCLEO-F401RE 開發(fā)板。

IIC配置

串口配置

開啟X-CUBE-ALS軟件包

時(shí)鐘樹配置

ADC使用定時(shí)器觸發(fā)采樣

在app_als_adc_utils.c中，定義了ADC使用的頻率，為8000Hz。

定時(shí)器的arr設(shè)置為10500-1，那么定時(shí)器頻率為8000Hz。
Trigger Event Selection ：update event 定時(shí)器自動(dòng)更新。

配置ADC檢測(cè)VD6283TX的GPIO2管腳的AD值。
設(shè)置觸發(fā)方式為外部觸發(fā)，選擇剛剛配置的TIM2，觸發(fā)方式為上升沿觸發(fā)。

開啟中斷。

KEIL配置

FFT代碼配置

arm_cortexM4lf_math.lib 庫(kù)包含了一系列數(shù)學(xué)函數(shù)，特別是適用于基于Cortex-M4和Cortex-M7處理器的浮點(diǎn)運(yùn)算單元的優(yōu)化數(shù)學(xué)例程。這些例程涵蓋了常見的數(shù)學(xué)運(yùn)算，如信號(hào)處理、濾波、變換等。
arm_math.h 這個(gè)頭文件包含了CMSIS-DSP庫(kù)的函數(shù)聲明、宏定義和結(jié)構(gòu)體定義等，可以通過包含這個(gè)頭文件，使用庫(kù)中提供的各種數(shù)學(xué)函數(shù)，包括信號(hào)處理、濾波、變換等。
添加arm_cortexM4lf_math.lib文件。

同時(shí)導(dǎo)入arm_math.h文件。

app_x-cube-als.c

由于需要進(jìn)行FFT算法，所以需要添加對(duì)應(yīng)數(shù)學(xué)頭文件。

#define ARM_MATH_CM4
#include "arm_math.h"
#include "app_als_adc_utils.h"

添加對(duì)應(yīng)的函數(shù)申明。

#define FLK_CHANNEL	(5U)

/*
 * Increasing the value of the FLK_DATA_SIZE symbol will increase
 * processing time, flicker accuracy and memory footprint
 */
#define FLK_DATA_SIZE (1024U)
#define FFT_SIZE (FLK_DATA_SIZE)

/* Private variables ---------------------------------------------------------*/
static uint8_t is_quit_requested;
static uint8_t is_autogain_requested;
static int16_t flk_data[FLK_DATA_SIZE];
volatile uint8_t ALS_EventDetected;

/*
 * The FFT of a real N-point sequence has even symmetry in the frequency domain.
 * The second half of the data equals the conjugate of the first half flipped in frequency.
 * Looking at the data, we see that we can uniquely represent the FFT using only N/2 complex numbers.
 * These are packed into the output array in alternating real and imaginary components:
 * X = { real[0], imag[0], real[1], imag[1], real[2], imag[2] ... real[(N/2)-1], imag[(N/2)-1 }
 */
static arm_rfft_fast_instance_f32 instance_fft;
static float32_t fft_in[FLK_DATA_SIZE];
static float32_t fft_out_tmp[FFT_SIZE];
static float32_t fft_out[FFT_SIZE/2];




/*
 * The FFT of a real N-point sequence has even symmetry in the frequency domain.
 * The second half of the data equals the conjugate of the first half flipped in frequency.
 * Looking at the data, we see that we can uniquely represent the FFT using only N/2 complex numbers.
 * These are packed into the output array in alternating real and imaginary components:
 * X = { real[0], imag[0], real[1], imag[1], real[2], imag[2] ... real[(N/2)-1], imag[(N/2)-1 }
 */
static arm_rfft_fast_instance_f32 instance_fft;
 
 
static void MX_VD6283A1_AnalogFlicker_Process(void); 

static float32_t complex_abs(float32_t real, float32_t complex);
static void init_fft(arm_rfft_fast_instance_f32 *instance, uint32_t size);
static void perform_fft(arm_rfft_fast_instance_f32 *instance, int16_t *data, float32_t *ffti, float32_t *ffto, uint32_t size);
static void find_flk_freq(uint32_t fs, float32_t *ffto, uint32_t *freq, uint8_t skip_dc, uint32_t size);


static int32_t flicker_autogain(uint8_t Instance, uint32_t *pAppliedGain, uint32_t timeoutMs);


static void display_gain(uint32_t gain);

在MX_VD6283A1_LuxCCT_Init()函數(shù)中添加init_fft快速傅里葉變換初始化。

static void MX_VD6283A1_LuxCCT_Init(void)
{
  /* Initialize Virtual COM Port */
  BSP_COM_Init(COM1);

  printf("VD6283TX Lux / CCT Examplenn");
  display_commands_banner();

    /* initialize ARM FFT library */
    init_fft(&instance_fft, FFT_SIZE);
	
  status = VD6283A1_LIGHT_SENSOR_Init(LIGHT_SENSOR_INSTANCE_0);

  if (status)
  {
    printf("VD6283A1_LIGHT_SENSOR_Init failedn");
    while(1);
  }
}

初始化完畢之后，添加頻率獲取函數(shù)。

static void MX_VD6283A1_AnalogFlicker_Process(void)
{
    uint32_t fs; /* sampling frequency */
    uint32_t pos = 0; 
    uint32_t flk_freq = 0;
    uint32_t index;

    uint32_t current_gain;
    uint32_t current_exposure;

    /* initialize exposure time */
    VD6283A1_LIGHT_SENSOR_SetExposureTime(LIGHT_SENSOR_INSTANCE_0, 100000);
    VD6283A1_LIGHT_SENSOR_GetExposureTime(LIGHT_SENSOR_INSTANCE_0, ¤t_exposure);
    printf("Exposure set to %lu usn", (unsigned long)current_exposure);

    /* initialize gain */
    flicker_autogain(LIGHT_SENSOR_INSTANCE_0, ¤t_gain, 1);
    printf("Channel %u gain set to", FLK_CHANNEL);
    display_gain(current_gain);

    status = als_adc_start(&fs);

    if (status)
    {
        printf("ADC Start failedn");
        while (1);
    }

    VD6283A1_LIGHT_SENSOR_StartFlicker(LIGHT_SENSOR_INSTANCE_0, FLK_CHANNEL, LIGHT_SENSOR_FLICKER_ANALOG);

    while (!is_quit_requested)
    {
        status = als_adc_get_frame(&flk_data[pos], &index);

        /* fill the ADC frame buffer */
        if (status == 0)
        {
            pos += ADC_FRAME_SIZE;
        }

        /* if the ADC frame buffer is full, then process it */
        if (pos == FLK_DATA_SIZE)
        {
            perform_fft(&instance_fft, flk_data, fft_in, fft_out, FFT_SIZE);
            find_flk_freq(fs, fft_out, &flk_freq, 1, FFT_SIZE);
            pos = 0; /* reset position index */

            printf("Flicker freq: %4lu Hzr", (unsigned long)flk_freq);
            fflush(stdout);

            if (is_autogain_requested == 1)
            {
              VD6283A1_LIGHT_SENSOR_StopFlicker(LIGHT_SENSOR_INSTANCE_0);

              flicker_autogain(LIGHT_SENSOR_INSTANCE_0, ¤t_gain, 1);
              printf("Channel %u gain set to", FLK_CHANNEL);
              display_gain(current_gain);

              VD6283A1_LIGHT_SENSOR_StartFlicker(LIGHT_SENSOR_INSTANCE_0, FLK_CHANNEL, LIGHT_SENSOR_FLICKER_ANALOG);
              is_autogain_requested = 0;
            }
        }

        handle_cmd(get_key());
    }

    als_adc_stop();

    VD6283A1_LIGHT_SENSOR_StopFlicker(LIGHT_SENSOR_INSTANCE_0);
    VD6283A1_LIGHT_SENSOR_DeInit(LIGHT_SENSOR_INSTANCE_0);

    printf("Quitting the demo...n");
    while (1);
}

在MX_X_CUBE_ALS_Process函數(shù)中開啟頻率獲取函數(shù)，關(guān)閉光強(qiáng)獲取函數(shù)MX_VD6283A1_LuxCCT_Process。

添加增益設(shè)置函數(shù)。

/*
 * @brief find and apply appropriate gain value depending on saturation value
 * @warning this function mustn't be called when a capture is ongoing
 */
static int32_t flicker_autogain(uint8_t Instance, uint32_t *pAppliedGain, uint32_t timeoutMs)
{
  int32_t res;
  uint8_t i, j;
  uint8_t idx = 7; /* start with mid-table value */
  const uint8_t sat_limit = 2;
  uint32_t saturation;

  /* duplicate 0x42AB to avoid 100x and keep multiples of 2 for array size */
  const uint16_t Gains[] = {
    0x42AB, 0x42AB, 0x3200, 0x2154, 0x1900, 0x10AB, 0x0A00, 0x0723,
    0x0500, 0x0354, 0x0280, 0x01AB, 0x0140, 0x0100, 0x00D4, 0x00B5
  };

  /* clip timeout value */
  timeoutMs = timeoutMs == 0 ? 1 : timeoutMs;
  timeoutMs = timeoutMs >= 100 ? 100 : timeoutMs;

  for (i = 0; i <= 3; i++)
  {
    VD6283A1_LIGHT_SENSOR_SetGain(Instance, FLK_CHANNEL, Gains[idx]);
    VD6283A1_LIGHT_SENSOR_GetGain(Instance, FLK_CHANNEL, pAppliedGain);

    res = VD6283A1_LIGHT_SENSOR_StartFlicker(Instance, FLK_CHANNEL, LIGHT_SENSOR_FLICKER_ANALOG);
    
    if (res)
      return res;

    /* read saturation value each ms so we can exit early if saturation detected */
    for (j = 0; j < timeoutMs; j++)
    {
      HAL_Delay(1);

      res = VD6283A1_LIGHT_SENSOR_GetSaturation(Instance, &saturation);

      if (res)
        return res;
      if (saturation > sat_limit)
        break;
    }

    res = VD6283A1_LIGHT_SENSOR_StopFlicker(Instance);
    
    if (res)
      return res;

    /* update index to next value */
    if (i)
      idx += saturation > sat_limit ? 1 < < (i - 1) : -(1 < < (i - 1));
    else if (saturation > sat_limit)
      idx++;
  }

  /* clip index if it reaches max value */
  if (idx > 15) 
    idx = 15;

  VD6283A1_LIGHT_SENSOR_SetGain(Instance, FLK_CHANNEL, Gains[idx]);
  res = VD6283A1_LIGHT_SENSOR_GetGain(Instance, FLK_CHANNEL, pAppliedGain);

  return res;
}

在下方添加函數(shù)的定義。

/*
 * @brief initilize arm rfft library
 */
static void init_fft(arm_rfft_fast_instance_f32 *instance, uint32_t size)
{
  arm_rfft_fast_init_f32(instance, size);
}

打印增益函數(shù)。

/*
 * @brief normalize, convert and dislay gain 
 */
static void display_gain(uint32_t gain)
{
    uint32_t g = (gain * 100) / 256;

    printf(" %3lu.%02lun", (unsigned long)g / 100, (unsigned long)(g % 100));
}

執(zhí)行FFT。

/*
 * @brief perform fft on the input buffer using arm rfft library
 */
static void perform_fft(arm_rfft_fast_instance_f32 *instance, int16_t *flk, float32_t *ffti, float32_t *ffto, uint32_t size)
{
    uint32_t i;
    uint32_t index = 0;

    /* copy the ADC sampled signal into the fft input buffer
     * this allows to convert the data from int16_t to float32_t */
    for (i = 0; i < size; i++)
    {
        ffti[i] = flk[i];
    }

    /* Perform the FFT on the input buffer:
     * results are packed in a way so that even indexes contain real values
     * and odd indexes contain the complex value of each bin.
     * Therefore the fft_output array contains FFT_SIZE / 2 bins */
    arm_rfft_fast_f32(instance, ffti, fft_out_tmp, 0);

    /* Calculate the magnitude for each bin from the temp fft output buffer */
    for (i = 0; i < size; i += 2)
    {
        ffto[index] = complex_abs(fft_out_tmp[i], fft_out_tmp[i+1]);
        if (ffto[index] < 0) ffto[index] = 0;
        index++;
    }
}

查找峰值頻率值。

/*
 * @brief find peak frequency value
 */
static void find_flk_freq(uint32_t fs, float32_t *ffto, uint32_t *freq, uint8_t skip_dc, uint32_t size)
{
    uint32_t i;
    uint32_t res;
    uint32_t index_max = 0;
    uint32_t limit = size / 2;

    float32_t max_value = -1;

    /* do not take account of the DC value if the flag skip_dc is set */
    skip_dc ? (i = 1) : (i = 0);

    /* run through the output array to detect the peak */
    for (; i < limit; i++)
    {
        if (ffto[i] > max_value)
        {
            index_max = i;
            max_value = ffto[i];
        }
    }

    /* convert index of the bin into frequency */
    res = (index_max * fs) / size;

    /* return the result if the pointer is valid */
    if (freq)
    {
        *freq = res;
    }
}

計(jì)算一個(gè)復(fù)數(shù)的絕對(duì)值。

/*
 * @brief compute absolute value of a complex number
 */
static float32_t complex_abs(float32_t real, float32_t complex)
{
  float32_t res;

  arm_sqrt_f32(real * real + complex * complex, &res);

  return res;
}

需要添加函數(shù)

arm_cortexM4lf_math.lib 庫(kù)包含了一系列數(shù)學(xué)函數(shù)，特別是適用于基于Cortex-M4和Cortex-M7處理器的浮點(diǎn)運(yùn)算單元的優(yōu)化數(shù)學(xué)例程。這些例程涵蓋了常見的數(shù)學(xué)運(yùn)算，如信號(hào)處理、濾波、變換等。
arm_math.h 這個(gè)頭文件包含了CMSIS-DSP庫(kù)的函數(shù)聲明、宏定義和結(jié)構(gòu)體定義等，可以通過包含這個(gè)頭文件，使用庫(kù)中提供的各種數(shù)學(xué)函數(shù)，包括信號(hào)處理、濾波、變換等。
app_als_adc_utils.c功能主要包括啟動(dòng)和停止ADC采樣，獲取采樣數(shù)據(jù)，ADC采樣速度設(shè)置，以及處理相關(guān)的硬件中斷。
app_als_adc_utils.h是app_als_adc_utils.c對(duì)應(yīng)頭文件。

審核編輯：湯梓紅