How to work with the desktop screen using robotjs and opencv4nodejs?

I used one of the examples. I want to take a screenshot of the desktop.
And find on it the picture I'm uploading. The problem is that I do not understand how to correctly convey the screenshot of the desktop. I am getting a gray window and an incorrect image.

```
const cv = require('opencv4nodejs');
const robot = require('robotjs');
var Jimp = require( 'jimp' );

const findWaldo = async () => {

var screen = robot.screen.capture(0, 0, 1024, 768); // I take screenshot here

const originalMat = new cv.Mat(screen.image, 1024, 768, cv.CV_8UC3); //Here's the processing

const waldoMat = await cv.imreadAsync(`${__dirname}/bobber.png`);
const matched = originalMat.matchTemplate(waldoMat, 5);

const minMax = matched.minMaxLoc();
const { maxLoc: { x, y } } = minMax;

// Draw bounding rectangle
originalMat.drawRectangle(
new cv.Rect(x, y, waldoMat.cols, waldoMat.rows),
new cv.Vec(0, 255, 0),
1,
cv.LINE_8
);

cv.imshow('We\'ve found Waldo!', originalMat);
cv.waitKey();

};
```
I am getting such a gray image. [https://i.stack.imgur.com/7AJ4F.png](https://i.stack.imgur.com/7AJ4F.png)

How can a full-screenshot of desktop image be obtained for recognition. In this case, do not save the file but work with the buffer?