Self-hosted personal gallery application with online photo management and albums, supporting Live/Motion Photos, EXIF parsing, geolocation recognition, and an explore map.
TypeScript albumsexif-extractiongeocodinglive-photomotion-photophoto-galleryphoto-sharingwebgl-viewer
import type { MediaLogger } from './logger.ts'
import sharp, { type Sharp } from 'sharp'
export interface HistogramData {
r: number[]
g: number[]
b: number[]
gray: number[]
}
/**
* 使用现有的 Sharp 实例计算图片的直方图数据
* @param sharpInst Sharp 实例
* @returns 直方图数据,包含 RGB 和灰度通道的 256 个值
*/
export const calculateHistogram = async (
sharpInst: Sharp,
logger?: MediaLogger,
): Promise<HistogramData> => {
try {
// 获取原始像素数据,确保是 RGB 格式
const { data, info } = await sharpInst
.clone() // 克隆 Sharp 实例以避免影响原实例
.toColourspace('srgb')
.removeAlpha() // 移除 alpha 通道
.raw()
.toBuffer({ resolveWithObject: true })
const { width, height, channels } = info
const totalPixels = width * height
// 初始化直方图数组(256个bin,对应0-255的像素值)
const histogramR = Array<number>(256).fill(0)
const histogramG = Array<number>(256).fill(0)
const histogramB = Array<number>(256).fill(0)
const histogramGray = Array<number>(256).fill(0)
// 遍历像素数据计算直方图
for (let i = 0; i < data.length; i += channels) {
const r = data.readUInt8(i)
const g = data.readUInt8(i + 1)
const b = data.readUInt8(i + 2)
// 累计 RGB 通道直方图
histogramR[r] = (histogramR[r] ?? 0) + 1
histogramG[g] = (histogramG[g] ?? 0) + 1
histogramB[b] = (histogramB[b] ?? 0) + 1
// 计算灰度值(使用 ITU-R BT.709 标准)
const gray = Math.round(0.299 * r + 0.587 * g + 0.114 * b)
histogramGray[gray] = (histogramGray[gray] ?? 0) + 1
}
// 归一化直方图(转换为百分比)
const normalizeHistogram = (hist: number[]) => {
return hist.map((count) => count / totalPixels)
}
return {
r: normalizeHistogram(histogramR),
g: normalizeHistogram(histogramG),
b: normalizeHistogram(histogramB),
gray: normalizeHistogram(histogramGray),
}
} catch (error) {
logger?.error('Failed to calculate histogram', error)
throw error
}
}
/**
* 使用图片缓冲区计算直方图数据
* @param buffer 图片缓冲区
* @returns 直方图数据
*/
export const calculateHistogramFromBuffer = async (
buffer: Buffer,
logger?: MediaLogger,
): Promise<HistogramData> => {
const sharpInst = sharp(buffer)
return calculateHistogram(sharpInst, logger)
}