Missing module functions with manually built OpenCV

[Cruaier]

Hi there,

I am currently working on a Lambda layer for object detection trying to use this module. Unfortunately I am having some issues after installing it.

My build process of OpenCV:

curl -L https://github.com/opencv/opencv/archive/refs/tags/4.10.0.tar.gz | tar zx && \ cd opencv-4.10.0 && mkdir -p build && cd build && cmake \ -DCMAKE_INSTALL_PREFIX=/opt \ -DBUILD_LIST=dnn,objdetect,imgcodecs \ -DBUILD_TESTS=OFF \ -DBUILD_PERF_TESTS=OFF \ .. && cmake --build . --target install

I am setting the needed environment variables accordingly and then install the module via npm without issues. After installation I run some code to verify the installation:

import cv from '@u4/opencv4nodejs';
console.log(cv);
const image = cv.imread('image.jpg');

Which errors:

const image = cv.imread('image.jpg');
                 ^

TypeError: cv.imread is not a function
    at Object.<anonymous> (/var/task/test.cjs:4:18)
    at Module._compile (node:internal/modules/cjs/loader:1369:14)
    at Module._extensions..js (node:internal/modules/cjs/loader:1427:10)
    at Module.load (node:internal/modules/cjs/loader:1206:32)
    at Module._load (node:internal/modules/cjs/loader:1022:12)
    at Function.executeUserEntryPoint [as runMain] (node:internal/modules/run_main:135:12)
    at node:internal/main/run_main_module:28:49

Node.js v20.12.0

Logging the cv object I can see there are many functions, e.g. from dnn available...but not from imgcodecs.

Having a look into the lib64 directory of OpenCV I can see that imgcodecs lib is present.

Any leads on what is missing would be appreciated. Thanks.

[cristinawithout]

Doing the same thing. Exact same issue. Find any solution?

I've been debugging and found that imgcodecs module is missing when imported from node. I verified that it's present in my opencv, but no matter what I set for opencv4nodejs, that module doesn't get included.

[cristinawithout]

I found an alternative. Since only imgcodecs was missing, I just couldn't use imread or imwrite. But since I was already using Sharp & have a layer for it, I'm using it to convert the image instead.

# replace imread
async function loadImageWithSharpToMat(imagePath) {
    const { data, info } = await sharp(imagePath)
        .raw()
        .ensureAlpha()
        .toBuffer({ resolveWithObject: true });

    // RGBA → Mat
    const matRGBA = new cv.Mat(data, info.height, info.width, cv.CV_8UC4);

    // Convert to BGR (OpenCV default)
    const matBGR = matRGBA.cvtColor(cv.COLOR_RGBA2BGR);

    return matBGR;
}

function matToSharpBuffer(mat): Buffer {
    const matRGBA = mat.cvtColor(cv.COLOR_BGR2RGBA);
    return Buffer.from(matRGBA.getData());
}

# replace imwrite
async function saveMatAsImageSharp(imagePath: string, mat) {
    const buffer = matToSharpBuffer(mat);

    await sharp(buffer, {
        raw: {
            width: mat.cols,
            height: mat.rows,
            channels: 4, // RGBA
        }
    }).toFile(imagePath);
}