Question

我想评估复选框是否已从扫描图像中选中。我找到了像node-dv和node-fv这样的节点模块。但是在安装时我在mac上遇到以下错误。

../deps/opencv/modules/core/src/arithm1.cpp:444:51: error: constant expression evaluates to 4294967295 which cannot be narrowed to type 'int' [-Wc++11-narrowing]
static int CV_DECL_ALIGNED(16) v64f_absmask[] = { 0xffffffff, 0x7fffffff, 0xffffffff, 0x7fffffff };
                                                  ^~~~~~~~~~
../deps/opencv/modules/core/src/arithm1.cpp:444:51: note: insert an explicit cast to silence this issue
static int CV_DECL_ALIGNED(16) v64f_absmask[] = { 0xffffffff, 0x7fffffff, 0xffffffff, 0x7fffffff };
                                                  ^~~~~~~~~~
                                                  static_cast<int>( )
../deps/opencv/modules/core/src/arithm1.cpp:444:75: error: constant expression evaluates to 4294967295 which cannot be narrowed to type 'int' [-Wc++11-narrowing]
static int CV_DECL_ALIGNED(16) v64f_absmask[] = { 0xffffffff, 0x7fffffff, 0xffffffff, 0x7fffffff };
                                                                          ^~~~~~~~~~
../deps/opencv/modules/core/src/arithm1.cpp:444:75: note: insert an explicit cast to silence this issue
static int CV_DECL_ALIGNED(16) v64f_absmask[] = { 0xffffffff, 0x7fffffff, 0xffffffff, 0x7fffffff };
                                                                          ^~~~~~~~~~
                                                                          static_cast<int>( )
2 errors generated.
make: *** [Release/obj.target/libopencv/deps/opencv/modules/core/src/arithm1.o] Error 1
gyp ERR! build error 
gyp ERR! stack Error: `make` failed with exit code: 2
gyp ERR! stack     at ChildProcess.onExit (/Users/entapzian/.nvm/versions/node/v4.3.1/lib/node_modules/npm/node_modules/node-gyp/lib/build.js:270:23)
gyp ERR! stack     at emitTwo (events.js:87:13)
gyp ERR! stack     at ChildProcess.emit (events.js:172:7)
gyp ERR! stack     at Process.ChildProcess._handle.onexit (internal/child_process.js:200:12)

上述依赖是否是我问题的最佳解决方案？如果没有，请建议我一个很好的解决方案。

Answer 1

对不起延迟回答，昨天和今天我一直很忙。这是一个抓取图像预定义区域并确定复选框是填充还是空的示例。这只是一个起点，可能会有很大的改进，但如果扫描的图像具有良好的质量，它应该可以工作。

第一步是获取图像的像素。接下来，通过根据模式抓取它们，可以获得图像中包含复选框的区域。最后，通过比较图像中该区域的平均亮度与未选中框的基线亮度来评估是否选中该复选框。

我建议使用get-pixels Node.js包来获取图像像素。

以下是您可以调整以满足您需求的示例：

var get_pixels = require(‘get-pixels’);
var image_uri = 'path_to_image';

get_pixels(image_uri, process_image);

var pattern_width = 800, // Width of your pattern image
    pattern_height = 1100; // Height of your pattern image

// The pattern image doesn't need to be loaded, you just need to use its dimensions to reference the checkbox regions below
// This is only for scaling purposes in the event that the scanned image is of a higher or lower resolution than what you used as a pattern.

var checkboxes = [
    {x1: 10, y1: 10, x2: 30, y2: 30}, // Top left and bottom right corners of the region containing the checkbox
    {x1: 10, y1: 60, x2: 30, y2: 80}
];

// You'll need to get these by running this on an unchecked form and logging out the adjusted_average of the regions
var baseline_average = ??, // The average brightness of an unchecked region
    darkness_tolerance = ??; // The offset below which the box is still considered unchecked

function process_image(err, pixels) {

    if (!err) {

        var regions = get_regions(pixels);

        var checkbox_states = evaluate_regions(regions);

        // Whatever you want to do with the determined states

    }else{
        console.log(err);
        return;
    }

}

function get_regions(pixels) {

    var regions = [], // Array to hold the pixel data from selected regions
        img_width = pixels.shape[0], // Get the width of the image being processed
        img_height = pixels.shape[1], // Get the height
        scale_x = img_width / pattern_width, // Get the width scale difference between pattern and image (for different resolution scans)
        scale_y = img_height / pattern_height; // Get the height scale difference

    for (var i = 0; i < checkboxes.length; i++) {

        var start_x = Math.round(checkboxes[i].x1 * scale_x),
            start_y = Math.round(checkboxes[i].y1 * scale_y),
            end_x = Math.round(checkboxes[i].x2 * scale_x),
            end_y = Math.round(checkboxes[i].y2 * scale_y),
            region = [];

        for (var y = start_y; y <= end_y; y++) {
            for (var x = start_x; y <= end_x; x++) {
                region.push(
                    pixels.get(x, y, 0), // Red channel
                    pixels.get(x, y, 1), // Green channel
                    pixels.get(x, y, 2), // Blue channel
                    pixels.get(x, y, 3) // Alpha channel
                );
            }
        }

        regions.push(region);

    }

    return regions;

}

function evaluate_regions(regions) {

    var states = [];

    for (var i = 0; i < regions.length; i++) {

        var brightest_value = 0,
            darkest_value = 255,
            total = 0;

        for (var j = 0; j < regions[i].length; j+=4) {

            var brightness = (regions[i][j] + regions[i][j + 1] + regions[i][j + 2]) / 3; // Pixel brightness
            if (brightness > brightest_value) brightest_value = brightness;
            if (brightness < darkest_value) darkest_value = brightness;
            total += brightness;

        }

        var adjusted_average = (total / (regions[i].length / 4)) - darkest_value; // Adjust contrast
        var checked = baseline_average - adjusted_average > darkness_tolerance ? true : false;

        states.push(checked);

    }

    return states;

}

从node.js中的扫描图像评估复选框

1 个答案: