Java calls opencv for face cropping, and found that when uploading a face image of about 8m, each face image cropping process consumes about 5-6g of memory, causing the program to easily crash. The following is the optimized code. And specify the memory at startup
nohup java -XX:MetaspaceSize=256m -XX:MaxMetaspaceSize=256m -Xms8192m -Xmx8192m -Xmn3072m -jar /dit/management-client.jar > /dit/management.log &You can use google thumbnailator to compress images.
package com.ahies.dit.management.service;
import cn.hutool.core.img.ImgUtil;
import cn.hutool.log.LogFactory;
import com.ahies.dit.management.model.File;
import com.google.common.collect.Maps;
import lombok.NonNull;
import net.coobird.thumbnailator.Thumbnails;
import org.apache.commons.io.FileUtils;
import org.apache.commons.io.FilenameUtils;
import org.opencv.core.*;
import org.opencv.imgcodecs.Imgcodecs;
import org.opencv.imgproc.Imgproc;
import org.opencv.objdetect.CascadeClassifier;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframew